a cost and implications assessment...• ontario’s weather would put businesses at a competitive...

Renewables-Based Distributed Energy Resources in Ontario

A Cost and Implications Assessment

Final Report

Marc Brouillette

June 2018

Renewables DER in Ontario – Cost & Implications Assessment

i

Final Report – June 2018

Executive Summary

Ontario’s Long-Term Energy Plan (LTEP) places significant reliance upon Distributed Energy Resources

(DER) consisting of renewables and storage to address a gap in the province’s electricity supply mix that

will emerge over the next fifteen years. This report examines the economics of meeting this supply gap

with renewables-based (solar and wind) DER and how the intermittent output of these variable generation

sources interacts with storage to undermine those economics.

Background

Today DER is at the center of energy policy discussions around the world. DER can include solar panels,

electricity storage, small natural gas-fuelled generators, and controllable loads such as electric vehicles

and water heaters. These resources are typically connected to distribution networks and are smaller in

scale than traditional transmission grid-connected generation facilities that serve most of Ontario’s

demand1. DER that includes renewables coupled with storage is advocated as the low-cost, low-emission

supply alternative to fossil fuels and the basis for adding more intermittent renewables to the supply mix.

Three factors have played a critical role in this transition: renewables such as wind and solar are now

integral parts of the energy mix in many jurisdictions; the next generation of these technologies have

experienced dramatic cost declines; and, Lithium-ion (Li-ion) batteries for energy storage are following a

similar cost reduction path.

DER is recognized as a critical component in the evolution of smart grid innovations. Information and

communication technology (ICT) innovations in smart control technologies aim to facilitate the integration

of renewables and storage technologies and enable two new paradigms: (1) a new class of customers:

consumers and producers of electricity (“prosumers”); and (2) community-based microgrids and virtual

power plants. DER connected in a microgrid configuration has the potential to provide dispatch flexibility

at the local distribution level that natural gas-fired generation currently provides for the transmission grid.

Ontario’s LTEP identifies a growing capacity gap in the province’s electricity supply. With the retirement

of the Pickering Nuclear Generating Station and the subsequent gradual expiration of Ontario’s contracted

renewables and natural gas-fired generation, 30% of Ontario’s generation capacity will have to be

renewed or replaced by 20352, even under the LTEP’s low demand forecast.

The LTEP places significant emphasis and reliance upon renewables-based DER to address this supply gap

and recognizes that storage is required to mitigate the effects of the intermittent output from wind and

solar generation. The LTEP is focused on increasing the adoption of renewables-based DER to achieve

several benefits that are enabled by storage: utilities can defer or avoid “wires” investments through

“non-wires” DER solutions; and customers can generate, store and sell their own power, and ensure their

own reliable electricity, both during times of peak demand and during power outages. Yet, the degree to

which the variable nature of wind and solar generation impacts the ability of storage to provide these

benefits is not well understood.

1 IESO website 2 MoE, LTEP, 2017


ii


Context

Studies undertaken in Ontario have examined how DER could be used to mitigate electricity system

challenges resulting from the extensive deployment of renewables over the last decade.3 In contrast, this

report examines the economics of using wind-based DER or solar-based DER solutions to fill the emerging

supply gap by 2035, explicitly considering the impact of intermittency.

To deliver the expected benefits, DER must supply either the daytime energy demand profile or the

baseload demand profile required by Ontario’s future electricity system. To this end, three DER

configurations are examined:

1. Solar-based DER at the community level or microgrid-scale that integrates solar panels and Li-ion

battery storage;

2. Wind-based DER consisting of grid-connected wind farms integrated with adjacent compressed

air energy storage systems (CAES); and

3. Baseload-supplied distributed energy storage (DES) is an alternative approach that consists of

grid-connected baseload generation such as nuclear or hydro that supplies community level

distributed storage. This latter baseload-supplied DES option represents a pathway to a low

carbon economy that has received little attention in the climate change and DER discourse.

Findings

The unfortunate truth is that renewables-based DER solutions are not a cost-effective way to meet

Ontario’s electricity needs in 2035 because the intermittency of renewables output negatively impacts

the cost of storage. These intermittency costs outweigh the forecast cost declines of the renewables and

storage technologies. When storage assets are coupled with intermittent renewables, storage operations

focus on managing the intermittency of the renewables. When storage is coupled with a baseload supply,

storage operations can be focused on managing demand fluctuations, a more direct use of the capabilities

of storage. This study has produced the following three major findings:

1. Ontario’s Weather-induced intermittency undermines the economics of renewables-based DER

• Intermittency creates a need for gas backup, which leads to high Levelized Cost of Energy

(LCOE) from DER systems - LCOE of solar-based DER is $270/MWh, 10% higher than today.

• Full renewables-based DER rollout would add $0.7B/year (solar DER) to $3.4B/year (wind DER)

to Ontario’s total cost of electricity - an increase of 3% to 15% over the LTEP forecast.

• Small-scale residential renewables-based DER will remain too costly for several decades.

2. Ontario renewables-based DER would have a systemic 35% higher cost structure than the U.S.

• Ontario’s weather would put businesses at a competitive disadvantage on energy costs.

3. Ontario has a better option with nuclear baseload-supplied DES

• Nuclear-supplied DES LCOE of $160/MWh is ~60% of the solar-based DER LCOE.

• Nuclear-supplied DES could reduce Ontario’s electricity cost by over $2B/year, $5.5B/year less

than wind-based DER, and 20% less than U.S. – a competitive advantage for Ontario.

• Small modular reactors (SMRs) and carbon capture may be the lowest cost solutions.

3 IESO, Energy Storage, 2016; Essex, 2017


iii


Discussion of Findings

Finding #1 – Ontario’s Weather-induced intermittency undermines economics of renewables-based DER

Figure ES-1 summarizes the costs of the assessed DER options for meeting Ontario’s 2035 supply gap.4

The costs are compared to those of the existing system5 as well as other new developing technologies in

terms of the Levelized Cost of Energy (LCOE) and the costs expected from each generation type as required

to meet Ontario’s anticipated LTEP supply gap in 2035. Intermittency creates a need for gas backup, which

leads to a high LCOE from renewables-based DER systems.

1. LCOE of solar-based DER is $270/MWh, 20% higher than the $240/MWh cost of today’s supplies that

it would replace

a. Today’s supply mix would have an LCOE of $240/MWh (if include carbon pricing).

b. Solar-based DER would have an LCOE of $270/MWh in 2017 dollars6, or $301/MWh if

microturbines are deployed instead of combined cycle gas turbines (CCGT).

c. The LCOE of the wind-based DER option would be approximately $380/MWh.

4 The scenarios all assume the LTEP demand for 2035 within a system built on Ontario’s committed hydro and nuclear assets and reflect industry projected 2030 costs. 5 Existing system costs from OEB RPP, OPO 2015 embedded generation, IESO 2016 Year End data 6 All currencies in this document are in $2017 CAD except in Section 4.0 or where otherwise specified.

Figure ES-1 – Total Annual Cost of DER and LCOE Comparison, Ontario, 2030


iv


2. A full rollout of renewables-based DER could add $0.7B/year to $3.4B/year to Ontario’s cost of

electricity.

a. This is equivalent to a cost increase of 3% to 15% over the 2030 forecast LTEP costs7.

b. The DER/DES options have two distinct cost components: the cost of generation and storage; and

the cost of the backup natural gas-fired generation and peaking supply.

c. The future generation and storage cost of a solar-based DER option is projected to be $4.8B/year,

1.9 times the cost of a baseload-supplied DES system comprised of conventional nuclear

generation and Li-ion battery storage. The generation and storage costs for the wind-based DER

option are projected to be $7.4B/year, 2.9 times the cost of the nuclear baseload-supplied DES

solution.

d. All options include the same need for peaking natural gas-fired generation plants to satisfy the

extreme demand peaks that occur on a few days every summer. The cost for 3,000 MW of peaking

gas supply in 2030 is forecast to be about $380M/year.8

Natural gas-fired generation would still be required to supply 20% to 30% of the incremental

daytime demand mostly as a result of seasonal variations in both demand and generation. The

estimated future share of natural gas-fired generation output could range from 3% to 5% of the

Ontario supply mix, similar to the 4% in 2017, but less than the 8% realized in 20169.

The cost of the backup natural gas-fired generation required by wind-based DER is $1.9B/year,

12% more than the $1.7B/year required for solar-based DER and 62% more than the $1.2B/year

required for the nuclear baseload-supplied DES option. The wind-based DER cost is higher due to

a much greater need for backup gas-fired generation capacity.

Some proposed DER schemes involve the use of microturbines in lieu of the grid-based

generation. The incremental cost of a microturbine was examined for the solar-based DER option.

Microturbines would increase the cost by 9% due to higher capital costs, low capacity factor and

carbon pricing.

3. The impact of renewables intermittency on the LCOE of DER/DES options in Ontario is illustrated in

Figure ES-2. Intermittency results in excess unutilized generation, conversion losses in the storage

system, low capacity factors of the storage asset, and the need for backup generation.

a. The LCOE of the solar-based DER has four contributing components:

• The cost of solar panels is based on the forecast LCOE for grid-connected solar of US

$47/MWh (for low cost areas of the U.S. with high levels of sunshine10). That same technology

installed at community-scale in Ontario will cost $120/MWh, a generation premium of

$73/MWh.

7 It is assumed that the OPO Outlook B total cost forecast of $20.2B/year in 2030 is the basis for the LTEP. 8 EIA 2017 Annual Energy Outlook, Strapolec analysis. 9 IESO Year End Data, 2016, 2017 10 Lazard LCOE v11, 2017


v


• Solar output intermittency combined with demand fluctuations increases the cost of storage

and solar output by $91/MWh for the energy that is used.

• DER solutions do not eliminate surplus from intermittent renewable energy production. Up

to 30% of solar energy will be curtailed or lost through storage inefficiencies – 19% of wind.

• Natural gas will be required to backup up the solar energy and supply 30% of the demand

increasing the total LCOE by $6/MWh to $270/MWh.

b. Wind-based DER solutions are costlier at $380/MWh.

4. Residential renewables-based DER will be uneconomic for decades

To best provide the desired system asset optimization and customer benefits, DER solutions should

be located as close to the demand load as possible, preferably on the consumers’ premises.

Unfortunately, as DER systems are moved closer to loads, the scale of the DER installation decreases:

a 1.5 MW solar panel could supply a community of 1,000 homes and businesses; for a single home, a

0.25 kW solar panel could provide all of the daytime energy that would be needed above what could

be supplied by Ontario’s committed baseload. The components of 5 kW or smaller scale DER solutions

are prohibitively expensive without the substantial subsidies that have promoted their use.

a. Cost forecasts show residential solar-based DER solutions will remain uneconomic beyond

2030.

Figure ES-2 – Ontario vs U.S. DER LCOE Contributions


vi


b. For solar-based DER, community-scale solutions may be the most promising DER option.

Increasing the size of DER installations to grid-scale solar offers little system benefits or cost

advantage.

c. Wind-based DER is only economic when using grid-scale wind, which also offers the potential

advantage of being paired with lower cost storage, such as compressed air energy storage

(CAES). However, grid-scale wind does not provide the desired DER benefit of reducing the

required capacity of the transmission and distribution systems. These must accommodate the

backup natural gas-fired generation capacity which is not reduced.

Finding #2 – Ontario renewables-based DER would have a systemic 35% higher cost structure than the U.S.

Figure ES-2 shows that the cost impacts of intermittency in Ontario are greater than in the U.S. This is

primarily due to the nature of Ontario’s geography and weather conditions, which lower the capacity

factors of the renewables. The higher capacity factors in the U.S. result from less variability or

intermittency of the renewable generation output.

1. The LCOE of the U.S. solar-based DER would be $200/MWh. The $270/MWh LCOE of the solar-

based DER in Ontario (using equivalent DER components) is 35% higher.

2. Similarly, the LCOE of wind-based DER may be 12% more in Ontario compared to the U.S.

3. Pursuing nuclear baseload-supplied DES options in Ontario could create a 20% cost advantage

over the U.S. solar-based DER options.

Finding #3 – Of the known and proven technology options, nuclear baseload-supplied DES will be the

lowest-cost option in any geography that has high renewables intermittency. Nuclear baseload-supplied

DES also has the greatest potential in Ontario to achieve the desired DER benefits of mitigating distribution

and transmission costs. Cost forecasts for other technology being developed suggest that:

1. The baseload-supplied DES would have an LCOE of $160/MWh. This option could reduce Ontario’s

annual electricity cost by over $2B.

2. Small modular reactors (SMRs) may be the lower cost solution for a broad range of jurisdictions

and locations compared to conventional nuclear;

3. Natural gas-fired generation (CCGT shown) equipped with carbon capture and sequestration (CCS)

may also be a low-cost low-carbon generation option.11 However, CCGT with CCS would not offer

the cost benefits from distribution and transmission system asset optimization and would not be

emission-free – three times more emissions than the solar-based DER and four times the

emissions of the nuclear baseload-supplied DES.

11 Assuming an operating capacity factor of 49%.


vii


Implications

1. Renewables-based DER should not be looked to as a cost-effective solution for Ontario’s emerging

capacity gap identified in the 2017 LTEP.

• Renewables-based DER can only be justified today based on either direct subsidies or indirect

subsidies enabled by market arbitrage.

• Investment in residential scale renewables-based DER is uneconomic. Incentives such as net-

metering, noted in the LTEP, are an indirect subsidy that would increase the total cost of the

entire electricity system.

2. Ontario’s emerging capacity gap can be best addressed by procurement of up to 5,000 MW of new

low-emission baseload electricity supply by 2035. New baseload capacity is required in Ontario to

supply two needs:

a) To fill Ontario’s emerging capacity gap for baseload supply requires the procurement of over

2,250 MW of new low-emission baseload supply.

• These resources will be required as soon as possible after the Pickering Nuclear Generation

Station retires in 2024.

• Based on 2030 cost projections, using renewables-based DER to perform a baseload function

will cost three to four times more than new nuclear stations and will not be emissions-free

because of the requirement for backup natural gas-fired generation. With the cost projections

predicated on significant cost declines to 2030, procurements prior to 2030, such as to replace

the retiring Pickering Nuclear Generation Station, will be costlier.

b) To meet daytime demand, another 2,700 MW of low-emission baseload supply is required by

2035 in order to implement the low-cost baseload-supplied DES.

3. Given the immediate requirement for low-emission baseload generation to fill the emerging capacity

gap after 2024, planning for such procurement should begin as soon as possible to best advance a

potential Ontario competitive cost advantage with respect to the U.S.

Analysis of the transmission, distribution, and reserve capacity benefits should be conducted. It would

likely show improved relative economics of baseload-supplied DES. Such analyses could additionally

inform policy and investment decision-makers about the economics of DER/DES solutions, their ability to

help address Ontario’s emerging capacity gap, and the potential to reduce overall electricity system costs.


viii


Table of Contents

Executive Summary ........................................................................................................................................ i

1.0 Introduction ...................................................................................................................................... 1

1.1 Background ................................................................................................................................... 2

1.2 Structure of this Document .......................................................................................................... 4

2.0 Methodology ..................................................................................................................................... 5

3.0 Distributed Energy Resources Capabilities and Applications............................................................ 7

3.1 The DER Promise ........................................................................................................................... 7

3.2 Locational Considerations for DER .............................................................................................. 10

3.3 Future Demand for DER Output Capabilities .............................................................................. 13

3.4 Candidate DER Technologies ...................................................................................................... 14

3.5 DER Scenarios .............................................................................................................................. 17

3.6 Summary ..................................................................................................................................... 18

4.0 Understanding the Cost of Renewables + Storage ......................................................................... 19

4.1 Renewable Generation Costs ...................................................................................................... 22

4.2 Storage Costs .............................................................................................................................. 28

4.3 Integrating Renewables with Storage ......................................................................................... 34

4.4 Case Examples of DER Economics ............................................................................................... 40

4.5 Comparison with Conventional Generation ............................................................................... 47

4.6 Summary of Cost Implications .................................................................................................... 52

5.0 Reality of Intermittency .................................................................................................................. 54

5.1 Intermittency Defined and Characterized with Demand ............................................................ 55

5.2 Solar-Based DER Intermittency Implications .............................................................................. 59

5.3 Wind-Based DER Implications ..................................................................................................... 65

5.4 Demand Fluctuation Implications ............................................................................................... 75

5.5 Summary Implications: Comparative Performance of DER Options........................................... 89

6.0 Ontario’s Geographical Challenge .................................................................................................. 93

6.1 DER Component LCOEs for Ontario ............................................................................................ 94

6.2 Total System Costs .................................................................................................................... 101

6.3 Intermittency and Demand Fluctuation Impacts on LCOE ....................................................... 103

6.4 U.S. Geography Induced Competitive Disadvantages for DER ................................................. 111

6.5 Summary of Ontario’s Geographic Cost Disadvantage ............................................................. 116

7.0 Summary Findings and Observations............................................................................................ 118


ix


Acknowledgements ................................................................................................................................... 123

Appendix A – References and Bibliography .............................................................................................. 124

Appendix B – List of Abbreviations ........................................................................................................... 126

Contact Information .................................................................................................................................. 128


x


List of Figures

Figure 1 – Ontario’s Peak Supply and Demand Outlook ............................................................................... 2

Figure 2 – Blended Cost of Ontario’s Energy Supply, 2016 ........................................................................... 3

Figure 3 – DER’s Promised Benefits to the Electricity System and Consumers ............................................ 8

Figure 4 – DER Potential for Peak Shaving .................................................................................................. 10

Figure 5 – DER Opportunities by Location .................................................................................................. 11

Figure 6 – Ontario’s Capacity Gap – Supply Needed for Baseload and Peak Demand Profiles .................. 14

Figure 7 – Wind Turbine System Illustrated ................................................................................................ 16

Figure 8 – DER vs. Conventional Generation Average U.S. LCOE ............................................................... 19

Figure 9 – Storage U.S. Average LCOS, 2017 vs. 2030 ................................................................................ 20

Figure 10 – Integrated DER LCOE Comparison ........................................................................................... 21

Figure 11 – Solar U.S. Capital Cost Forecast, 2017 vs. 2030 ....................................................................... 23

Figure 12 – Wind U.S. Capital Cost Forecast, 2017 vs. 2030 ....................................................................... 23

Figure 13 – Solar and Wind Average U.S. LCOE Forecast, 2017 vs. 2030 ................................................... 23

Figure 14 – Solar Community and Commercial Scale Capital Cost Forecast to 2040 ................................. 24

Figure 15 – Solar Community and Commercial Scale Capital Cost CAGR Forecast to 2040 ....................... 25

Figure 16 – Solar Community LCOE Forecast to 2030 ................................................................................ 26

Figure 17 – Grid-Scale Solar Capital Cost Forecast to 2030 ........................................................................ 26

Figure 18 – Grid-Scale Solar LCOE Forecast to 2030 ................................................................................... 27

Figure 19 – Wind Capital Cost Forecast to 2040 ......................................................................................... 27

Figure 20 – Storage U.S. Average Capital Cost, 2017 vs. 2030 ................................................................... 28

Figure 21 – Storage U.S. Average LCOS, 2017 vs. 2030 .............................................................................. 29

Figure 22 – Lithium-Ion Battery Physical Energy Storage System .............................................................. 30

Figure 23 – Lithium-Ion Battery Capital Cost CAGR Forecast to 2030 ........................................................ 31

Figure 24 – Lithium-Ion Battery Capital Cost Forecast to 2030 .................................................................. 31

Figure 25 – Illustrative System Costs: LCOS by Category ............................................................................ 32

Figure 26 – Lithium-Ion Battery LCOS CAGR Forecast to 2030 ................................................................... 33

Figure 27 – Lithium-Ion Battery LCOS Forecast to 2030 ............................................................................. 33

Figure 28 – Pumped Hydro and Compressed Air LCOS Forecast to 2030 ................................................... 34

Figure 29 – Integrated Solar Community DER LCOE ................................................................................... 35

Figure 30 – Integrated Wind + CAES LCOE .................................................................................................. 35

Figure 31 – Solar Community DER LCOE Components ............................................................................... 35

Figure 32 – Conceptual Arrangement of a DC Coupled PV + Storage Battery System ............................... 37

Figure 33 – Avoided Cost of Battery BOS Associated with PV Coupling ..................................................... 37

Figure 34 – Lithium-Ion Battery Capital Cost Reduction ............................................................................. 37

Figure 35 – Lithium-Ion Battery Integration LCOS Reduction ..................................................................... 37

Figure 36 – Solar and Storage Daytime Demand Supply ............................................................................ 38

Figure 37 – Solar and Storage for Baseload Supply .................................................................................... 38

Figure 38 – Wind + CAES LCOE Components .............................................................................................. 39

Figure 39 – Wind Capacity Sizing for Daytime Community Demand .......................................................... 40

Figure 40 – Wind Capacity Sizing for Baseload Demand ............................................................................ 40

Figure 41 – Residential Net-Metering with 4 kW Rooftop Solar PV ........................................................... 42

Figure 42 – Net Energy Metering ................................................................................................................ 43


xi


Figure 43 – Optimal Dispatch of a DC Tightly Coupled PV plus Storage System ........................................ 44

Figure 44 – Storage Charging During Low Cost Overnight Periods ............................................................. 45

Figure 45 – LCOE of Nuclear and Natural Gas-Fired Generation ................................................................ 48

Figure 46 – Nuclear LCOE Forecast to 2030 ................................................................................................ 49

Figure 47 – Combined Cycle Gas Turbine with CCS LCOE Forecast to 2030 ............................................... 49

Figure 48 – Nuclear Baseload-Supplied DES LCOE Components ................................................................ 50

Figure 49 – Nuclear and DES for Daytime Supply ....................................................................................... 51

Figure 50 – Natural Gas-Fired Generation LCOE Comparison .................................................................... 52

Figure 51 – Intermittency and Demand Fluctuation Impact on DER Component Use ............................... 54

Figure 52 – Relationship Between Wind, Solar, & Community Demand .................................................... 55

Figure 53 – Supply Mix Contributions to Ontario Demand ......................................................................... 57

Figure 54 – Components of Community Demand ...................................................................................... 58

Figure 55 – Daily Supply of Committed Resources ..................................................................................... 58

Figure 56 – Average Daily September Community Demand ...................................................................... 59

Figure 57 – Daily Solar Output Profile ......................................................................................................... 61

Figure 58 – Solar and Storage for Daytime Demand Supply ....................................................................... 62

Figure 59 – Solar and Storage System for September 2017 ....................................................................... 62

Figure 60 – Solar Capacity Factor Rolling 28-Day Average ......................................................................... 63

Figure 61 – Solar Output Comparison by Year ............................................................................................ 64

Figure 62 – Solar and Storage System with Unadjusted Gas Peak ............................................................. 65

Figure 63 – Solar and Storage System with Gas Averaged as Percent of Demand ..................................... 65

Figure 64 – Variations in Daily Wind Output .............................................................................................. 67

Figure 65 – Winter Wind Output as a % of Capacity .................................................................................. 68

Figure 66 – Hourly Wind Generation .......................................................................................................... 69

Figure 67 – Storage Size for Smoothing Wind Intermittency ..................................................................... 69

Figure 68 – Integrating Wind and Storage .................................................................................................. 70

Figure 69 – Cumulative Stored Wind Energy .............................................................................................. 70

Figure 70 – Maximum 28-Day Stored Energy Swing ................................................................................... 71

Figure 71 – Wind Output Comparison by Year ........................................................................................... 73

Figure 72 – Impact of Intermittency and Demand on DER ......................................................................... 75

Figure 73 – Annual Community Daytime Demand Profile .......................................................................... 76

Figure 74 – Average Daily Winter Community Daytime Demand Profile ................................................... 76

Figure 75 – Average Daily Summer Community Daytime Demand Profile ................................................. 76

Figure 76 – Daily March Demand Profile .................................................................................................... 77

Figure 77 – Daily May Demand Profile........................................................................................................ 77

Figure 78 – Daily September Demand Profile ............................................................................................. 77

Figure 79 – Community Daytime Weekly Demand Profile ......................................................................... 78

Figure 80 – Yearly Ontario Demand Profile Comparison ............................................................................ 78

Figure 81 – Solar and Storage System for September 2017 ....................................................................... 79

Figure 82 – Solar-based DER System Behavior ........................................................................................... 81

Figure 83 – Required Demand Storage Outflow ......................................................................................... 82

Figure 84 – Wind-based DER System Behavior ........................................................................................... 84

Figure 85 – Nuclear and Storage System for May....................................................................................... 85

Figure 86 – Nuclear Baseload-Supplied DES Behavior ................................................................................ 88


xii


Figure 87 – DER Generation Required for Community Daytime Demand .................................................. 90

Figure 88 – Intermittency Implications on LCOE of Ontario DER Options .................................................. 93

Figure 89 – Ontario vs. U.S. LCOE Comparison Supplying Daytime Demand ............................................. 94

Figure 90 – Solar (Community-Scale) Ontario LCOE Components .............................................................. 95

Figure 91 – Wind (Grid-Scale) Ontario LCOE Components ......................................................................... 95

Figure 92 – Electrical Regions in the U.S. .................................................................................................... 98

Figure 93 – Annual Hours of Sunshine ........................................................................................................ 99

Figure 94 – Solar (Grid-Scale) LCOE and Capacity Factor .......................................................................... 100

Figure 95 – Wind (Grid-Scale) LCOE and Capacity Factor ......................................................................... 101

Figure 96 – Total Annual Cost of DER and LCOE Comparison, Ontario, 2030 .......................................... 103

Figure 97 – Intermittency Implications on LCOE of Ontario DER Options ................................................ 104

Figure 98 – Community Solar-Based DER Component LCOE Contributions ............................................. 105

Figure 99 – Grid-Scale Wind + CAES Component LCOE Contributions ..................................................... 108

Figure 100 – Nuclear Baseload-Supplied DES Component LCOE Contributions ....................................... 110

Figure 101 – LCOE Comparison of DER Solutions ..................................................................................... 111

Figure 102 – Ontario vs U.S. DER LCOE Contributions .............................................................................. 112

Figure 103 – Solar Capacity Factor Comparison ....................................................................................... 112

Figure 104 – Solar Capacity Factor Implications on U.S. Cost Drivers ...................................................... 114

Figure 105 – Ontario vs. U.S. Solar-Based DER Cost Implications ............................................................. 114

Figure 106 – Average Monthly Wind Capacity Factor .............................................................................. 115

Figure 107 – Wind Capacity Factor Implications on U.S. Cost Drivers ...................................................... 116

Figure 108 – Ontario vs. U.S. Grid-Scale Wind-Based DER Cost Implications ........................................... 116

Figure 109 – Total Annual Cost of DER for Ontario, 2030......................................................................... 118

Figure 110 – Ontario vs U.S. DER LCOE Contributions .............................................................................. 120


xiii


List of Tables

Table 1 – DER Use Cases by Location .......................................................................................................... 11

Table 2 – Direct Benefits of Energy Storage ............................................................................................... 12

Table 3 – Ontario Energy Storage Project Summary .................................................................................. 15

Table 4 – Lithium-Ion Battery Capital Cost Components ............................................................................ 30

Table 5 – Types of Renewables Intermittency and Variable Generation ................................................... 56

Table 6 – Solar Constant Demand, System Characteristics ........................................................................ 60

Table 7 – Solar Constant Demand, System Performance ........................................................................... 60

Table 8 – Wind Constant Monthly Daytime Demand ................................................................................. 65

Table 9 – Wind Baseload Demand, System Characteristics ........................................................................ 65

Table 10 – Wind Constant Monthly Demand, System Performance .......................................................... 66

Table 11 – Wind Baseload Demand, System Performance ......................................................................... 66

Table 12 – Wind Constant Monthly Demand for Winter, System Characteristics ..................................... 72

Table 13 – Wind Constant Monthly Demand for Winter, System Performance ........................................ 72

Table 14 – Wind Constant Monthly Demand, 3 Year Comparison ............................................................. 74

Table 15 – Wind Baseload Demand, 3 Year Comparison ............................................................................ 74

Table 16 – Comparison of Solar Constant Demand and Full Year Demand Cases ...................................... 80

Table 17 – Wind Storage Size Sensitivity Analysis ...................................................................................... 83

Table 18 – Comparison of Nuclear Constant Demand and Full Year Demand Cases ................................. 86

Table 19 – Sensitivity Analysis Results for Nuclear-Based DER Storage Size .............................................. 86

Table 20 – Full Year Demand System Characteristics for DER Options ...................................................... 89

Table 21 – Full Year Demand Performance Metrics for DER Options ......................................................... 91

Table 22 – Imported Content (%)................................................................................................................ 96

Table 23 – Regional Cost Multiplication Factors ......................................................................................... 97

Table 24 – Ontario vs. U.S. LCOE of Generation ......................................................................................... 98

Table 25 – System Architecture Assumptions for DER Options ................................................................ 102

Table 26 – Supply Mix Contribution to Total Ontario Demand ................................................................ 102

Table 27 – Factors Impacting Solar-Based DER Costs ............................................................................... 105

Table 28 – Factors Impacting Wind-Based DER Costs ............................................................................... 108

Table 29 – Factors Impacting Nuclear-Based DER Costs ........................................................................... 109

Table 30 – U.S. Solar Capacity Adjustments ............................................................................................. 113

Table 31 – U.S. Wind Capacity Adjustments ............................................................................................. 115


1


1.0 Introduction

Today, Distributed Energy Resources (DER) are at the center of energy policy discussions around the

world. DER can include solar panels, electricity storage, small natural gas-fuelled generators, and

controllable loads such as electric vehicles and water heaters. These resources are typically connected to

distribution networks and are smaller in scale than traditional transmission (Tx) grid-connected

generation facilities that serve most of Ontario demand12. DER that includes renewables coupled with

storage are advocated as the low-cost, low-emission supply alternative to fossil fuels and the basis for

adding more intermittent renewables to the supply mix. Three factors have played a critical role in this

transition: renewables such as wind and solar are now integral parts of the energy mix in many

jurisdictions; the next generation of these technologies have experienced dramatic cost declines; and,

Lithium-ion (Li-ion) batteries for energy storage are following a similar cost reduction path.

DER is recognized as a critical component in the evolution of smart grid innovations. Information and

communication technology (ICT) innovations in smart control technologies aim to facilitate the integration

of renewables and storage technologies and enable two new paradigms: (1) a new class of customers:

consumers and producers of electricity (“prosumers”); and (2) community-based microgrids and virtual

power plants. DER connected in a microgrid configuration has the potential to provide dispatch flexibility

at the local distribution level that natural gas-fired generation currently provides for the Tx grid.

The LTEP is focused on increasing the adoption of renewables-based DER to achieve several benefits

enabled by storage:

“Renewable distributed generation can benefit local distribution companies (LDCs) and their customers:

Utilities can defer or avoid certain costly investments in their local distribution networks; and Customers

can generate and store their own power, lowering bills and ensuring reliable access to electricity when

power from their network is not available.”13

Several studies undertaken in Ontario have examined how DER could be used to mitigate the electricity

system challenges resulting from the extensive deployment of renewables over the last decade.14 In

contrast, this study looks at the emerging capacity gap and examines the economics of how new

renewables-based DER solutions could fill the supply gap by 2035, explicitly considering the impact of

intermittency.

To deliver the benefits expected, DER must supply either the daytime energy demand profile or the

baseload demand profile required by Ontario’s future electricity system. Baseload-supplied distributed

energy storage (DES) is an alternative approach that consists of grid-connected baseload generation such

as nuclear or hydro that supplies community level distributed storage. This latter baseload-supplied DES

option represents a pathway to a low carbon economy that has received little attention in the climate

change and DER discourse.

12 IESO website 13 MoE, LTEP, 2017, pg. 68 14 IESO, Energy Storage, 2016; Essex, 2017


2


The implications of DER scenarios are examined across four factors, with specific attention paid to the

impacts that renewables intermittency has on storage:

1. The total costs of renewables-based DER solutions in Ontario;

2. Potential for wind and solar renewables-based DER to fill the capacity gap and supply the

forecasted demand;

3. How those costs compare to baseload-supplied DES resources; and

4. How costs may differ between Ontario and the U.S.

1.1 Background

Ontario’s LTEP identifies a growing capacity gap in the province, as illustrated in Figure 115. By 2035, 30%

of Ontario’s generation capacity will have to be renewed or replaced16. The most significant declines in

available capacity occur in 2023/2025 and 2029/2030.

The change in the 2023/2025 timeframe is due to the retirement of the Pickering Nuclear Generating

Station (PNGS). This will remove approximately 20% of nuclear baseload supply from Ontario’s supply mix

creating a future need for a baseload supply replacement. It will also create the need for greater voltage

regulation services east of Toronto.17

The remaining contributions to the capacity gap, including the large change in 2029/2030, are due to the

gradual expiration of the20-year supply contracts for renewables and gas-fired generation. Much of this

supply addresses Ontario’s daytime demand profile and peak reserve requirements. This includes the

15 Figure extracted from the 2017 LTEP 16 One of the criticisms of the 2017 LTEP is that it did not address any demand expectations that may arise from emissions reduction. Should there be appreciable demand growth as has been forecasted by many, then additional capacity will be required beyond what is shown in Figure 1. 17 IESO, Energy Storage, 2016, p.29

Figure 1 – Ontario’s Peak Supply and Demand Outlook


3


expiration of the contract for Ontario’s only peaking gas plant capable of ramping up quickly enough to

meet Ontario’s flexible supply requirements.18

The existing and committed resources identified in Figure 1 that remain in 2035 are primarily low

emission, low cost assets and include: Ontario’s hydro fleet, refurbished nuclear, and biomass. Along with

the import/export energy exchange capability with Hydro Quebec, these resources provide Ontario with

a low-carbon, flexible baseload capability.

Figure 2 contrasts the average cost of the committed baseload resources at $66/MWh against the average

cost of the expiring assets at $223/MWh19. The expiring assets reflect high cost resources. Ontario has an

opportunity to switch out these high cost resources and replace them with lower cost options.

With respect to committed resources, nuclear refurbishment is the single largest asset renewal

component of the 2017 LTEP. The Financial Accountability Office of Ontario (FAO) has estimated that the

cost of the refurbished nuclear will be $80/MWh when the program is completed20, a 15% increase over

the cost of nuclear today. The cost of refurbished nuclear represents a benchmark for the total generation

cost to provide baseload power. It also offers a reference by which DER applications that would supply a

baseload equivalent can be measured for their impact on the cost of power assumptions in the LTEP. A

cost of $223/MWh is assumed for supplying the rest of Ontario’s demand requirements and is a

benchmark for comparing the cost-effectiveness of deploying DER solutions to supply the province’s

daytime demand.

The LTEP identifies three broad initiatives intended to help find lower cost options to replace the expiring

contracts:

1. Independent Electricity System Operator (IESO) Market Renewal Initiative

18 York Energy Center per IESO Report: Energy Storage, 2016, p.25 19 OEB RPP, 2017 20 FAO Nuclear Refurbishment, 2017

Figure 2 – Blended Cost of Ontario’s Energy Supply, 2016


4


2. LDC Grid Modernization

3. Integrated Regional Resource Plans (IRRP)

The latter two initiatives both support and enable renewables-based DER and instruct the Ontario Energy

Board (OEB) to develop regulations that price these options to accelerate their adoption.

The LTEP places significant emphasis and reliance upon renewables-based DER to address this supply gap

and recognizes that storage is required to mitigate the effects of the intermittent electricity production

from wind and solar generation. The LTEP is focussed on increasing the adoption of renewables-based

DER to achieve several benefits that are enabled by storage: utilities can defer or avoid “wires”

investments through “non-wires” DER solutions; and customers can generate, store and sell their own

power, and ensure their own reliable electricity, both during times of peak demand and during power

outages. Yet, the degree to which the variable nature of wind and solar generation impacts the ability of

storage to provide these benefits is not well understood. The focus of this report is to assess the ability of

renewables-based DER to cost-effectively address Ontario’s pending capacity gap.

1.2 Structure of this Document

The body of this report consists of six main sections.

Section 2.0 provides a brief overview of the methodology deployed in researching costs and modelling

Ontario’s future energy system.

Section 3.0 examines the DER promise of expected benefits, introduces the definitions of possible DER

locations, summarizes sample technologies that are being deployed, describes the demand conditions

that DER should be expected to solve, and defines what DER concepts are examined.

Section 4.0 addresses the perception in the public domain that renewables and storage costs are declining

so rapidly that they will be the best economic choice of all generation options. This section presents the

research findings that quantify what the expected future costs of renewables and storage technologies

are expected to be. The implications for integrating solar and wind technologies with storage in the DER

concept are described. The concepts that are currently being used to establish economic viability of

renewable-based DER are summarized. The expected costs for conventional technologies are also

presented.

Section 5.0 examines the nature of the intermittency of renewables and the associated implications on

the use of storage. The nature of demand fluctuations is also described along with the implications on the

use of storage for both renewables-based DER as well as baseload-supplied DES.

Section 6.0 interprets the U.S. dollar cost forecast for DER technologies and applies that to the Ontario

situation. The unique cost implications of renewables’ intermittency in Ontario on the DER options are

explored along with the impact of Ontario’s expected future demand fluctuations. The ability of the

DER/DES options to supply the needed future demand is summarized. The cost implications are compared

for the DER options for Ontario’s low-emission supply mix.

Finally, Section 7.0 summarizes the findings and offers several observations.


5


2.0 Methodology

The following steps were taken to establish the basis for the findings in this report:

1. Establishing the Ontario context for DER and the associated options.

2. Identifying the global consensus on future cost expectations for renewables and storage.

3. Analyzing the impact of the intermittency of renewables and demand fluctuations on Ontario’s

DER options.

Establishing Ontario’s DER context

The 2017 LTEP provides the foundational perspective for Ontario’s plans to leverage renewable-based

DER, including the forecast of Ontario’s capacity to be renewed and /or replaced. The IESO defined the

impacts of DER and storage challenges for the operation of the Tx system, and the Ontario Ministry of

Energy (MoE) subsequently commissioned a study on the potential benefits of storage for the province.

Forecasting Future Costs

Several sources were drawn upon to develop a consensus on the future costs of generation, renewables

and storage. The approach taken for this study was generally to accept the most aggressively low-cost

forecast for the renewables and storage available. The purpose of embedding a low-cost bias in this

study’s assumptions was to underscore the significance of the resulting high total system costs after taking

into account the impact of intermittency. This represents a conservative approach.

Generation costs were obtained from three U.S based sources published in 2017:

1. U.S. Energy Information Administration (EIA) 2017 Annual Energy Outlook (AEO)

• The EIA bases its forecasts on actual projects and applies an economy of scale function for

lessons learned approach to estimate future costs to 2050.

• The primary EIA reference was the “Levelized Cost and Levelized Avoided Cost of New

Generation Resources” in the 2017 AEO.

• The research also drew upon reports that the EIA commissioned to support its 2017 AEO

estimates for small solar and wind applications.

▪ Distributed Generation and Combined Heat & Power System Characteristics and

Costs in the Buildings Sector, 2017 (prepared by Leidos).

2. U.S. National Renewable Energy Laboratory (NREL) 2017 Annual Technology Baseline (ATB)

• The NREL mandate is to keep at the forefront of technology development in this space.

3. Lazard’s Levelized Cost of Energy Analysis – Version 11.0

• Lazard surveys developers of technologies for their expectations on costs, both recent and for

the next five years.

Storage cost forecasts were derived and validated based on four sources:

1. The primary source was Lazard’s Levelized Cost of Storage (LCOS) 3.0.

• The earlier version, LCOS 2.0, was also consulted, as it provided estimates for pumped storage

and compressed air energy storage (CAES).

• As Lazard only provides directional long-term forecasts, the forecast used in this study was

developed based on a review of industry commentaries, insights, and other proxies.


6


2. Several secondary NREL research reports were also consulted to understand component costs and

the opportunities that may arise when integrating solar with storage.

3. The results of the forecasts used here were then compared to forecasts from the 2018 Green Tech

Media (GTM)21 and the 2017 International Renewables Energy Agency (IRENA) reports to validate that

the 2030 values used in this report were also lower than forecasts from these sources.

Analysing Intermittency

To conduct the intermittency analysis, IESO data was obtained for the years 2015 through to 2017. This

data included:

1. Generation output by hour, for all Ontario generation including wind and solar.

• Wind and solar data was obtained for both before and after curtailment.

2. Demand data for 2015 through to 2017.

To assess the implications of demand fluctuations on DER solutions, the 2017 LTEP was consulted to define

incremental assumptions to be added to the IESO 2016 Ontario Planning Outlook (OPO) low growth

demand Outlook B scenario. From these sources, a detailed hourly forecast was developed for 2035.

Separate simulations were conducted for wind-based DER, solar-based DER, and nuclear baseload-

supplied DES options. Assumptions were made on how to size the storage for each case to best illustrate

the impacts of intermittency. High level sensitivity assessments were conducted to illustrate the impact

of storage capacity. There are many parameters that could be tuned to optimize an actual

implementation. However, the results described in this report suggest that fine tuning is not expected to

materially change the relative outcomes of the scenarios.

Benchmark data was obtained for U.S. jurisdictions from the EIA and Lazard to assess capacity factor

differences between the U.S. and Ontario. These capacity factor differences form the basis for projecting

relative impacts from intermittency.

21 GTM, 2018


7


3.0 Distributed Energy Resources Capabilities and Applications

DERs are being viewed as a potential game changer with respect to how future electricity systems are

planned and developed22. According to Ontario’s IESO23, DERs are:

“…electricity-producing resources or controllable loads that are directly connected to a local

distribution system or connected to a host facility within the local distribution system.

DERs can include solar panels, combined heat and power plants, electricity storage, small-natural

gas-fuelled generators, electric vehicles and controllable loads, such as heating/cooling systems

and electric water heaters. These resources are typically smaller in scale than the traditional

generation facilities that serve most of Ontario demand.”

DER that includes renewables coupled with storage are advocated as the low-cost, low-emission supply

alternative to fossil fuels and the basis for adding more intermittent renewables to the supply mix. Three

factors have played a critical role in this transition: renewables such as wind and solar are now integral

parts of the energy mix in many jurisdictions; the next generation of these technologies have experienced

dramatic cost declines; and, Lithium-ion (Li-ion) batteries for energy storage are following a similar cost

reduction path.

Many proponents of renewables-based DER also advocate that DER represents an alternative to nuclear.

Conventional power stations, such as coal-fired, gas and nuclear-powered plants, as well as hydroelectric

dams and large-scale solar power stations, are centralized and often require electric energy to be

transmitted over long distances. By contrast, DER systems are decentralized, modular and flexible

technologies, that are located close to the load they serve, typically with capacities of 10 megawatts (MW)

or less. These systems can comprise multiple generation and storage components. In contrast, many

believe that nuclear must play a significant role in reducing emissions from the production of energy, in

particular for 24x7 baseload supply. The hyperbole around DER falls into two categories:

1. The degree to which distributed storage can be coupled with renewables to mitigate

intermittency and enable more renewables; and,

2. The degree to which DER can provide broad system benefits beyond just smoothing intermittency.

This section summarizes the DER promise, the types of DER installations that have been identified in the

literature, the demand profiles that DER should respond to, and examples of DER technologies that have

been considered. This section concludes with a definition of the DER options that are contrasted in the

costing and intermittency assessments of this report.

3.1 The DER Promise

The DER promise is that distributed energy production, coupled with appropriate amounts of storage, can

provide several benefits to both the electricity system and to consumers, as summarized in Figure 324.

22 IESO Energy Storage, 2016; MoE, LTEP, 2017 23 IESO website 24 Siemens 2011; LTEP 2017; CPI 2017; Massachusetts Energy Storage Initiative, 2016; Lazard LCOS v3.0; Mowat 2017


8


System benefits are mostly enabled by the capabilities of storage and fall into two main categories:

1. Increased Grid Flexibility, Resiliency and Reliability

• Among the sought-after benefits for grid reliability, where high renewables are deployed or

desired, is the potential for distributed storage to allow for easier integration of intermittent

renewable generation.

o According to the IESO25, energy storage can be used to enhance the grid’s ability to

manage the influx of variable renewable generation in the following areas: load

following, ramping and dispatch flexibility; regulation; Tx voltage control; operating

reserve; and zonal limitations. The IESO also recognizes that storage is not the only

available option for addressing these operating challenges.

• Decentralizing energy generation and storage would also leave the grid less vulnerable to local

or system-wide disruptions or centralized cyber attacks and also supports quicker local

energizing of the distribution system following a blackout26. However, energizing the higher

voltage Tx system requires substantial reactive power support be available on-line at both

ends of long Tx lines suggesting DER solutions, may have limited benefit.

• Distributing storage and generation resources throughout the grid, among and closer to user

demand centres should allow for better response to localized demand fluctuations.

2. Increased Asset Capacity Factors for Generation, Tx, and Distribution (Dx)

• DER has the potential to provide a local demand management function that can smooth out

the magnitude of electricity demand peaks that are imposed upon the grid. By doing so, DER

25 IESO Energy Storage, 2016 26 Essex Energy, 2017

Figure 3 – DER’s Promised Benefits to the Electricity System and Consumers


9


can reduce the capacity requirements for the grid and hence, if the same energy is delivered,

increase the capacity factors of generation, Tx and potentially Dx assets in the bulk electricity

system.

• Increasing the capacity factor of the delivery systems reduces the fixed capacity costs and the

effective per MWh electricity rates.

• By potentially reducing the peak reserve margin requirements, DER can reduce the cost of

reserve capacity that is very infrequently used.

• The ability to more efficiently size generation, Tx, and Dx assets to best meet energy demand

lowers the fixed costs of capacity. Installed capacity represents the lion’s share of costs in a

low emission electricity system because fuel costs are low or zero. Reducing capacity

improves the cost-effectiveness of the entire system.

Consumer benefits fall into three categories of interest:

1. Reducing Energy Bills Through System Benefits

• System benefits of DER impact consumers through their energy bills. The ability of DER to help

optimize system capacities will reduce the effective delivery infrastructure costs that are paid

for by the consumer.

• By locating generation closer to the consumer, electricity losses during delivery to consumers

would be reduced.

2. Generate Revenue for Customers with DERs

• For consumers that have rooftop solar panel based DER systems, experience shows that the

electricity output does not generally match the consumer’s energy use. The ability to sell the

excess energy back to the grid allows the consumer to recover some of the costs of the

installation. Net metering is one example of how such revenue opportunities can be enabled.

Unfortunately, most pricing incentives that promote the deployment of high cost residential

generation systems are subsidies that increase the costs for other ratepayers. This is discussed

in Section 4.4.

• Storage, on its own, creates the opportunity for consumers to charge the storage when retail

rates are low, such as at night, and then discharge the energy back to themselves or the

system when retail rates are higher. Effectively, consumers can participate in retail pricing

arbitrage and benefit from low cost electricity. The pricing arbitrage also offers benefits to

the system to the extent that peak demand shaving can be consistently achieved as illustrated

in Figure 427.

3. Emergency Backup Generation

• Some consumers feel the need to have an emergency backup electricity supply for when

unexpected outages occur on the bulk electricity system. Such backup capability can be

provided by the storage capability of DER. Since diesel backup generators are relatively costly

and environmentally unfriendly, this DER benefit may be important to those customers

wanting emergency backup. However, electrical storage would typically be configured with a

shorter backup run-time than diesel generators due to the higher cost.

27 Massachusetts Energy Storage Initiative, 2016


10


3.2 Locational Considerations for DER

DER can be deployed at many different points within the electricity system. As shown in Figure 528,

locational options fall into two categories: in front of the meter; and behind the meter. In either case,

installations can vary significantly in scale. Residential and commercial DER opportunities are referred to

as “behind the meter”, as they are installed on consumer premises and are not metered by the utility.

These installations are typically very small, with a capacity on the order of 1 to 5 kW for a single home but

could be up to 300 kW for commercial applications. In the case of storage, the amount of stored energy

typically provides only a few hours of rated capacity.

“In front of the meter” applications involve solutions that would typically be managed by the LDC. These

solutions range in size from smaller community installations to large-scale Dx installations. A small

community installation would typically serve 1,000 homes with 1 MW to 1.5 MW of generation capacity.

A larger Dx scale installation would be optimized around the Dx substations to help smooth peak demands

and would serve larger communities of 10,000 homes for example.

The largest scale resources are high voltage grid connected resources typically in excess of 30 MW, which,

by definition, are not usually considered to be DER.

28 Definitions adapted from Lazard LCOS v3.0

Figure 4 – DER Potential for Peak Shaving


11


System benefits from DER vary with location as summarized in Table 129.

Table 1 – DER Use Cases by Location

DER Location Description and Use

Fro

nt-

of-

Met

er

Community Supports community or small power systems that can have some independence from the broader power grid. Could also provide ramping support to enhance system stability and increase reliability of service.

Distribution Typically placed at substations or distribution feeders controlled by utilities to defer distribution upgrades. May also provide flexible peaking capacity and mitigate stability problems.

Grid Large-scale energy system designed to replace peaking gas turbine facilities and the reliability services they provide. Can be brought online quickly to meet rapidly ramping demand for power at peak and taken offline quickly as power demand diminishes.

Beh

ind

th

e M

eter

Residential Behind-the-meter residential use to provide backup power and extend the usefulness of self-generation (e.g., “solar plus storage”). Smooths the quantity of electricity sold back to the grid from distributed solar PV applications.

Commercial Behind-the-meter peak shaving and demand charge reduction services for commercial energy users. Option to provide grid services to the utility.

Table 1 – DER Use Cases by Location

29 Paraphrased from Lazard LCOS v3.0

Figure 5 – DER Opportunities by Location


12


The Ontario MoE commissioned a study30 to evaluate the economics of using storage based on the value

elements of the DER promise. The study focused on the near-term value of DER for Ontario’s current

extensive deployment of renewables and the associated system challenges identified by the IESO. The

possible value of DER that could be obtained from different types of installations at different locations is

summarized in Table 231.

Table 2 – Direct Benefits of Energy Storage

Distributed Connected Energy Storage Location

Lazard/IRENA Framework Grid Distribution Community Residential/

Commercial

Benefits

Category Currently Monetizable Benefits

At Tx

Station

Middle of

Feeder

End of

Feeder

Behind

Meter

System

Non-Spinning Reserve Availability ✓ ✗ ✗ ✗

Spinning Reserve Availability ✓ ✗ ✗ ✗

Reserve Activation ✓ ✗ ✗ ✗

Power Quality Improvement ✓ ✓ ✓ ✓

Frequency Regulation ✓ ✓ ✓ ✗

Voltage Control ✗ ✓ ✓ ✗

Black Start ✓ ✓ ✗ ✗

Asset

Optimization

Distribution System Upgrade Avoidance ✓ ✓ ✓ ✗

New Generation Capacity Avoidance ✓ ✓ ✓ ✗

Reduce Dispatching of Peaker Facilities ✓ ✓ ✓ ✓

Consumer Price

Wholesale Market Arbitrage ✓ ✓ ✓ ✗

Retail Market Arbitrage ✗ ✗ ✗ ✓

Global Adjustment Charge Reduction (Class A) ✗ ✗ ✗ ✗

Consumer Redundant Power Supply (Reliability) ✓ ✓ ✓ ✗ Table 2 – Direct Benefits of Energy Storage

Essex’s report identified that grid-scale applications located at the Tx/Dx Interface offer the most benefits.

This is because grid-based wind intermittency reliability issues occur at this interface. Since 85% of the

installed wind in Ontario is connected to the Tx system, mitigating the impacts of variable generation is

best managed, as close to the generation source, as possible. From a Dx system perspective, the Tx/Dx

interface is the connection to the variable generation. The reverse is true for solar, which has been

predominantly installed at individual residences in Ontario. Essex’s approach to addressing these

challenges with DER is defined here as “intermittency management”.

30 Essex Energy, 2017 31 Recreated from Essex 2017 report to regroup rows into the broader categories used in this report.


13


In contrast, the approach taken in this report looks at how to best meet demand through renewables-

based DER. From that perspective, the concept of renewables-based DER entails co-locating the storage

with the generation, such as with behind the meter applications. Co-location is inherently designed to

mitigate the intermittency effects of the generation, but also delivers a broader “demand management”

function by ensuring the DER produces an output to meet consumer demand in a specific area. Effective

consumer demand management should flatten the load presented back to the Dx and Tx systems,

achieving the maximum asset optimization benefit.

To maximize the demand management benefits of DER and optimize system capacity factors, solutions

are best located as close as possible to the user demand. For demand management purposes, the

following expectations arise:

• Grid based solutions are likely to offer little Tx and Dx benefit as the purpose of grid-based DER

would be to mimic the capability of gas-fired generation plants to meet the demand on the grid.

Some avoided line loss benefits may be realized if their grid locations are closer to the demand

centres than heritage generation sites.

• Dx scale installations would smooth out grid demand but offer little benefit to the Dx networks

downstream where peak requirements remain driven by consumer behavior.

• When DER is implemented at the community level or behind the meter, the demand management

or levelling function is better positioned to maximally optimize the entire delivery system

infrastructure.

3.3 Future Demand for DER Output Capabilities

Many of the benefits of DER for smoothing peak demand to optimize delivery infrastructure have been

articulated. However, in order to meet the capacity gap challenges that are emerging in Ontario, DER

solutions should be expected to supply one of two components of user energy demand:

1. Daytime demand to mitigate the need to renew or replace expired gas plant operations.

2. Baseload demand to replace the 3,000 MW gap when the PNGS retires in 2024.

The two profiles, illustrated in Figure 6, reflect the requirements for a DER system that best supports

Ontario’s electricity system needs and yields the desired benefits.

Since renewable outputs are determined by the natural energy supply (wind, sun or water) and are only

dispatchable to the extent that the wind is blowing, the sun shining and water is available, these resources

are to a significant extent uncontrollable. These means storage systems are required to provide the output

that meets energy user demands across the entire electricity system.


14


To provide a baseload function, storage must smooth intermittent renewables at a lower cost than low

emission baseload generations such as nuclear or hydro, or even natural gas with carbon capture and

storage.

For renewables-based DER solutions to supply daytime demand, the requisite storage for managing

intermittency must be enhanced to also manage demand fluctuations. With a baseload-supplied DER

solution, the storage function must only manage the demand fluctuations. Section 5 discusses the nature

of intermittent renewables and quantifies the implications of supplying the required demand.

3.4 Candidate DER Technologies

Several storage technology demonstrations have been deployed in the marketplace. These provide

examples for the concepts modelled in this report:

• Li-ion batteries coupled with solar to emulate a natural gas peaker system

• Pumped hydro storage coupled with either hydro or wind resources

• CAES coupled with wind resources

A summary of storage technologies that are being piloted in Ontario is provided in Table 332.

Table 3 – Ontario Energy Storage Project Summary

Project Technology Capacity Benefits

POWER.HOUSE Li-ion Battery 228 kWh Redundant power supply

Penetanguishene Microgrid Battery 500 kWh Redundant power supply

Pan Am Games 2015 100 kVA, 125kWh Load shifting

32 Reproduced from Essex Energy, 2017

Figure 6 – Ontario’s Capacity Gap – Supply Needed for Baseload and Peak Demand Profiles


15


Hydrostor - Toronto CAES Varies Distribution line decongestion

eCAMION – Toronto Hydro Li-ion battery 25 kW, 16 kWh Infrastructure support

eCAMION - Toronto Li-ion battery 500 kW, 250 kWh Infrastructure support

NRStor - Minto Flywheel ±2 MW, 500 kWh Frequency regulation

HONI – Clear Creek Flywheel ±5 MW, 500 kWh Voltage control

Opus One - DEMSN Battery Voltage support, generation

integration

RES Canada - Strathroy Li-ion battery 4 MW, 2.6 MWh Frequency regulation

NEDO – Oshawa Li-ion battery 10 kWh Load levelling

Convergent Energy – Sault Ste. Marie

Li-ion battery 7 MW Reliability

Hydrogenics Power-to-Gas 2 MW Frequency regulation

Ameresco – Phase II Solid Battery (2x) 2 MW, 8 MWh Peak shaving

Baseload Power – Phase II Flow Battery 2 MW, 8 MWh Grid support and arbitrage

NextEra – Phase II Solid Battery 2 MW, 8 MWh Grid support and arbitrage

NRStor Inc. – Phase II CAES 1.75 MW, 7 MWh Grid support

SunEdison – Phase II Flow Battery 1 MW, 4 MWh

(2x) 2 MW, 8 MWh Grid support

Table 3 – Ontario Energy Storage Project Summary

Examples of underlying DER concepts and applications include:

a) Arizona’s Experience with Solar and Storage

One of the major benefits of pairing batteries with solar power is to fix the mismatch between patterns

of solar generation and demand. The demand peaks of the morning and evening in Arizona bookend

midday peak solar generation. Arizona, at times, has too much solar generation and has little need for

additional generation with the same profile.

Tucson Electric signed a power purchase agreement (PPA) with First Solar for a solar plus battery peaker

system in February 201833. The peaker system has been contracted to provide up to 50 MW of power

between 3pm and 8pm. The cost of this system is expected to be competitive with existing gas peaker

plants when completed in 2021. The storage system will be paired with a new 65 MW solar plant and be

able to store 135 MWh for almost 3 hours of discharge duration.

b) Ontario Power Generation (OPG) Niagara Falls Pumped Storage

33 GreenTech, 2018


16


OPG has a pumped hydro station next to the Sir Adam Beck hydro power complex, both of which are

adjacent to the Niagara River. Water is pumped into a reservoir at night when demand and electricity

prices are low. Filling the reservoir can take up to 8 hours. During high demand, water can be released to

flow through turbines at the Sir Adam Beck Complex providing up to 174 MW of capacity.34 The station

has a unique cascade configuration made possible by the unique geography of Niagara Falls and the

Niagara Escarpment. The effective power swing is several times greater than the power capacity of the

pumped storage facility.

c) Germany Integrated Wind Farm and Pumped Hydro Storage35

Large wind turbines in Gaildorf, Germany are connected to a nearby hydro pumped storage facility and

store water within the towers themselves as illustrated in Figure 7.

Water from the turbines can be released and channeled into a hydro power station below when needed.

Each of four 180-meter-high, 3.4 MW wind turbines can store up to 70 MWh of water pumped up from a

nearby lake. Stored water is equivalent to over 20 hours of full capacity wind generation. With wind blades

of over 60 Meters, these large grid-scale facilities are almost 80 stories high, larger than most skyscraper

office towers in the world’s major cities.

d) Toronto Hydro Compressed Air Energy Storage

Toronto Hydro is working with Hydrostor Inc. to analyze the electrical grid benefits of underwater CAES36.

The world's first system has now been installed in Lake Ontario. The pilot project will focus on the unit’s

ability to provide reserve power, shift load and smooth out Tx and Dx congestion. The system is designed

to store excess electricity generated during low-demand off-peak hours by driving compressed air into

34 OPG, 2018 35 Dvorak, 2017 36 Toronto Hydro, 2018

Figure 7 – Wind Turbine System Illustrated


17


a flexible wall air accumulator below the lake's surface. When the energy is required, the system is

reversed.

e) NRStor-Hydrostor Goderich CAES Demonstration Project37

This energy-storage facility with a four-hour discharge capability would provide energy to help the Ontario

electricity grid meet peak demand. The expected value for ratepayers is more efficient integrated wind-

generated electricity and load levelling. Grid operators need flexible resources to offset errors in wind and

solar forecasting.

The 1.75 MW, 7 MWh CAES facility will use proprietary technology to store electricity in the form of

compressed air and heat. NRStor’s novel solution utilizes an existing bedded salt cavern as a storage

cavity. The project is expected to demonstrate the success of fuel-free CAES, creating market

opportunities for Canadian companies to globally deploy locally developed technology.

3.5 DER Scenarios

The scenarios chosen for this study are:

1. Solar-based DER: Community-scale solar integrated with Li-ion battery storage

• This is one of the most prevalent architectures discussed in the literature.

• Stakeholder interviews conducted by the Essex study identified the 1 MW community scale

as the most likely form of distributed storage.38

• This option offers the best opportunity for optimising system capacity benefits and has

storage costs comparable to larger scale installations.

2. Wind-based DER: Grid-based wind coupled with CAES

• For the majority of jurisdictions, only grid-scale wind is expected to be a viable economic

option (see Section 4.0).39 Small-scale wind has relatively excessive land use implications,

even at small scale.40 The planned MoE’s net metering siting restrictions will effectively make

small scale wind ineligible for Ontario’s residential net metering program.41

• CAES is a lower cost grid-scale storage option than Li-ion batteries. Ontario’s topography and

resource extraction legacy may present many options for integrating CAES with wind farm

output. However, because of the thermodynamics of compressing air, the round-trip energy

losses of the process are less efficient than the Li-ion batteries.

3. Baseload-supplied DES: Grid-based nuclear baseload coupled with distributed Li-ion battery storage

systems

• Li-ion storage systems have the ability to be ubiquitously distributed at the community level

and hence, from a storage perspective, this scenario is analogous to solar.

37 Sustainable Development Technology Canada, 2018 38 Essex Energy, 2017 39 Remote community applications have not been considered 40 Leidos, 2016 41 Environmental Registry Regulation Proposal #013-1916, Proposed New Regulation to be made under the

Electricity Act, 1998 (28 November 2017).


18


3.6 Summary

There are many benefits that various DER configurations could provide to the electricity system and

consumers. However, not all of the consumer benefits are in the best interests of the overall electricity

system. The ability to realize the potential benefits also varies by location.

The system benefits from DER may be best achieved by designing DER solutions that leverage the energy

advantages of Ontario’s existing electricity systems.


19


4.0 Understanding the Cost of Renewables + Storage

The declining costs of renewables and energy storage solutions has received a lot of media coverage, with

advocates claiming that these technologies are now competitive with fossil fuels. Yet, the MoE-sponsored

study found that based on what can be monetized by investors in today’s markets, the storage options

are not economically viable, with the exception of commercial applications that aim to reduce demand

charges42.

This section explores the costs of renewables, storage and other generation required to enable DER

options. Specific attention is given to how the cost of renewables and storage may decline by 2030 and

how that compares to the other components of low emission DER system options.

The major findings for the costs of generation, storage, and integrated DER systems are illustrated below

in Figures 8, 9, and 10.

a) Costs of renewables are declining modestly

Figure 843 summarizes the expected future costs in the U.S., in terms of the Levelized Cost of Electricity

(LCOE) of the low emission generation options that could support the long-term objectives of DER: solar,

wind, nuclear and Combined Cycle Gas Turbine (CCGT) natural gas-fired generation with carbon capture

and sequestration (CCS). Community scale solar systems are expected to have material annual decline

rates with total declines of approximately 30% from 2017 to 2030. However, the LCOE for community-

scale solar installations will remain above $70/MWh44 in regions with average U.S. capacity factors.

42 Essex identified monetizable and non-monetizable value areas. Strapolec disagrees with how the non-monetizable benefits have been calculated, the assumption of 100% capacity factor of the storage, and the assumed ongoing existence of surplus energy. Essex use cases only economic due to pricing mechanisms arbitrage. Refer to section 4.4. 43 CAGR is compounded annual growth rate from 2017 to 2030; Natural gas plant assumes a 30% duty cycle for supplying daytime demand. Renewables costs without storage and subject to from intermittency, 44 All dollar figures in Section 4.0 are in US $2017 unless specified otherwise.

Figure 8 – DER vs. Conventional Generation Average U.S. LCOE


20


At this level, it would appear that the LCOEs of the grid-based solar and wind are indeed lower than

nuclear and natural gas with carbon capture. This section establishes that the LCOE of standalone

renewables is not the measure that should be used, but rather it is the LCOE of the integrated DER solution

that should be the key benchmark. For renewables to be viable in a DER context, they must supply a

particular demand load, such as the daytime load that exceeds baseload. To do so, the renewables must

be coupled with storage so that the energy provided is coincident with the demanded load. Power

engineers refer to this capability as capacity value. Without storage, intermittent renewables have

relatively little capacity value.

b) Costs of battery storage expected to decline 50% by 2030

Figure 945 summarizes the LCOS of storage systems evaluated in this study: Li-ion batteries of residential

to Dx scale; and, pumped hydro and CAES suitable for some grid-scale applications. At $363/MWh,

residential storage is expected to remain an excessively high cost option beyond the forecast time horizon

of this study. Even community scale storage is expected to remain higher than the more conventional

options of pumped hydro and CAES. The CAES technologies are expected to remain the least expensive

option by a wide margin looking forward to 2030.

The benefits of Li-ion batteries over the other options are their low loss factor and the flexibility to locate

them where they are needed. Pumped hydro and CAES are limited to where geological features or other

available physical characteristics enable their installation. They may have limited ability to support small-

scale DER solutions.

The raw costs of renewables and storage do not reflect the cost of a DER system. The cost of a full system

that integrates renewables with storage options to meet a demand requirement is the relevant measure

45 Storage costs shown do not include cost of energy to be stored and reflect operating duty cycles where storage is fully charged and discharged on a daily basis.

Figure 9 – Storage U.S. Average LCOS, 2017 vs. 2030


21


for assessing how competitive these technologies are. The integrated performance of the DER system is

determined by several factors:

1. Integrating solar and storage systems could allow for capital cost reductions due to sharing of

components.

2. Configuring the components (e.g. the solar panel and battery) to supply the expected demand

profile determines the blend of the electricity used.

3. Storage systems are not 100% efficient in the process of charging and discharging the energy to

and from the storage. This inefficiency is measured as the round-trip energy loss. Losing energy

through the storage device increases the net cost of energy that is output from the storage device.

• Li-ion batteries are expected to have a 14% round trip energy loss. CAES is expected to have

a 35% round trip energy loss.46

c) Costs of DER will remain high

Figure 1047 illustrates how the cost of renewable energy changes when coupled with storage. Three terms

capture how the costs are realized. The used generation is at the cost typically expected. Stored energy

has a higher cost for the generation that is stored because of the round-trip losses in the system. The cost

of the storage then gets added to that of the stored energy to get the full cost of the stored energy. The

net blended cost is a function of how much energy is used directly versus stored.

46 Lazard, LCOE v11, 2017 47 Solar case is community-scale, while the wind case is grid-scale with compressed gas storage. Values reflect ideal weather and demand conditions that do not introduce intermittency.

Figure 10 – Integrated DER LCOE Comparison


22


The storage required to help renewables deliver energy in response to user demand will add $107/MWh

to the cost of any stored solar for a total LCOE of stored energy of $194/MWh in the U.S. Storage costs

when integrated with solar photovoltaic (PV) panels in a tightly coupled manner can realize some capital

cost savings which have been reflected. The net blended solar DER cost of $134/MWh reflects a solution

where half of the solar energy is directed to storage for use in supplying the expected demand profile. At

$134/MWh, it is not clear whether solar-based DER is competitive with alternative generation.

For wind-based DER, the expected blended cost could be as low as $57/MWh. This would be for a grid-

scale application and could be competitive with other solutions. The blended cost of a nuclear baseload-

supplied DES solution is estimated at $120/MWh, 11% less than solar. Section 5 of this document explores

the implications intermittency has on the above LCOEs.

Given these results, why is so much attention being paid today to renewables-based DER? There are two

possible answers to this question: (1) DER is considered applicable for optimizing revenue capture in fossil-

based energy markets where prices peak with high demand; or, (2) creative subsidies, such as net

metering hide the full cost that is being incurred to the whole system.

As discussed earlier, this study examines the costs of available technologies to supply a full demand

profile.

This section examines the following relevant subjects:

1. Projected costs for solar and wind generation

2. Projected costs of storage

3. The cost implications for integrated DER solutions under ideal conditions

4. The marketing of DER solutions today

5. The costs of alternative low-emission technologies, such as nuclear and CCGT with CCS.

4.1 Renewable Generation Costs

This section examines the capital cost and LCOE forecasts of solar and wind technologies obtained from

multiple sources. For each technology, different scales of implementation are considered, from small-

scale residential applications to large grid-scale facilities. Figures 11, 12 and 13 illustrate the expected

capital cost and LCOE decline rates.


23


For solar applications, residential scale systems are expected to drop the fastest, but will remain 25%

higher than community solar and almost double that of grid-scale solar by 2030. The capital cost forecasts

for small-scale wind suggest it will remain 4-5 times the cost of grid-scale wind. Given land use challenges,

small-scale wind in a residential setting is not considered further in this study. Grid-scale wind is expected

to have more modest LCOE declines than solar.

4.1.1 Solar Cost Assumptions

Three different scales of solar installations were considered:

1. Residential installations, of 5 kW or less, which are rooftop mounted and feature no tracking

mechanisms;

2. Community and commercial scale installations, of up to 1.5 MW, consisting of both fixed-tilt

rooftop and pole mounted tracking systems; and,

171

113

6246

9172

47 37

$0

$50

$100

$150

$200

Solar Residential 5 kW

CF 16%

Solar Community 1.5 MW CF 22.5%

Solar Grid 100 MW CF 20%

Wind Grid 100 MW CF 41%

Solar and Wind Average U.S. LCOE, 2017 vs 2030($U.S./MWh, CAGR %)

2017 2030

4.8%

3.5%

2.1%

$/MWh

/yr

/yr

/yr

1.7% /yr

Solar Wind

Figure 11 – Solar U.S. Capital Cost Forecast, 2017 vs. 2030

Figure 12 – Wind U.S. Capital Cost Forecast, 2017 vs. 2030

Figure 13 – Solar and Wind Average U.S. LCOE Forecast, 2017 vs. 2030


24


3. Grid-scale solar installations from 5 MW to 150 MW.

Residential solar systems were not examined, as the capacity size for a single home solar panel is less than

1 kW and very expensive when connected to residential-scale storage.

Community Scale Solar Cost Forecasts

a) Capital costs

Commercial/community solar capital costs, in $/kW, have been projected by Leidos, Lazard, NREL, and

IESO48. The capacity of commercial installations varied from 300 kW (NREL) to 1 MW (Lazard). Lazard has

defined a community solar installation of 1.5 MW, which is the cost design case used in this study. The

community installation costing is based on an optimized fixed-tilt installation. Figure 14 shows in red the

estimated community-based solar installation capital cost. Most of the sources reflect costs for rooftop

commercial installations.

The average capital cost used for this analysis is $2,532/kW in 2017 and is expected to drop to $1,686/kW

by 2030.

Lazard’s 1.5 MW case is used to represent the 2017 solar capital costs for comparison with other types of

generation. The rationale is: (1) Lazard’s estimate is similar to the Leidos reference estimate; (2) The Leidos

system is the largest other system quoted; and (3) the IESO estimates could not be reconciled with the

others (after an assumed exchange rate discount of 15%).

48 Leidos, 2016; Lazard LCOE v11.0; NREL Annual Technology Baseline, 2017; IESO OPO, 2016

Figure 14 – Solar Community and Commercial Scale Capital Cost Forecast to 2040


25


The compound annual growth rates (CAGR) used for each of the projections are summarized in Figure 15.

Leidos provides low and high CAGRs for reference and advanced equipment respectively.

The average of the Leidos reference and advanced equipment CAGRs were chosen for the forecast in this

report yielding 4.1% for 2017-2020 and 3.2% for 2020-2030. These are near the high end of the rates of

decline in the sample set. Applying these CAGRs to the 2017 Lazard data creates a forecast for community-

scale solar with a capital forecast that declines from $2,532/kW in 2017 to $1,686/kW in 2030. This is

within the range of the IESO’s forecast.

b) LCOE Forecast

For this analysis, the LCOE is of greater interest and is what will be used directly in the cost comparisons.

Several sources provided estimates of the cost of solar in 2017, but only NREL developed a forecast. The

Strapolec forecast uses CAGRs from the above projected capital cost. This resulted in a lower cost than

that derived by NREL for 2030. This suggests the cost forecast for the DER option used in this report is

conservatively low.

4.1%3.4%

2.1%

0%

1%

2%

3%

4%

5%

6%

7%

2017-2020 2020-2025 2025-2030 2030-2035 2035-2040

Solar Community and Commercial Capital Cost CAGR Forecast to 2040(%)

IESO Commercial Rooftop (500 kW) IESO Small-Scale Ground-Mounted (500 kW)

Leidos Commercial Large - Reference (650 kW-DC) Leidos Commercial Large - Advanced (650 kW-DC)

NREL Commercial (300 kW) Strapolec Community Solar (1.5 MW)

2020-2030

3.2%

Figure 15 – Solar Community and Commercial Scale Capital Cost CAGR Forecast to 2040


26


Grid-Scale Solar Cost Forecasts

For grid-scale solar, Lazard, IESO, and NREL provide capital costs, as shown in Figure 17. NREL’s forecast

was 25% lower than the IESO’s forecast, which could reflect issues related to Ontario specific installations

(see Section 6.1). The CAGRs for grid-scale solar were smaller at only 1.3%/year after 2020 and lower again

after 2030. Nevertheless, grid-scale solar capital costs are expected to decline by 25% from 2017 to 2030.

Of the few LCOE forecasts found for this study, as shown in Figure 18, NREL’s forecast decline from

$62/MWh in 2017 to $47/MWh in 2030 was chosen to be conservative, as it demonstrates the highest

price decline. The capacity factor assumption for the solar installations is 20%.

113

140 141

75

113

72

$0

$20

$40

$60

$80

$100

$120

$140

$160

2017 2030

Solar Community LCOE Forecast to 2030($U.S./MWh)

Lazard Community (1.5 MW) Lazard Rooftop C&I (1 MW)

NREL Commercial (300 kW) Strapolec Community Solar (1.5 MW)

$/MWh

Figure 16 – Solar Community LCOE Forecast to 2030

Figure 17 – Grid-Scale Solar Capital Cost Forecast to 2030


27


4.1.2 Wind Cost Assumptions

Wind capital costs were projected by the EIA, Leidos, Lazard, and NREL. Figure 19 summarizes the capital

cost projections for a range of installations from these sources.

It is clear from Figure 19 that smaller scale wind installations are expected to remain four to six times as

costly as grid-scale applications. CAGRs are in the 1%/year range with the exception of one outlier, EIA’s

100 kW commercial application. This estimate is discounted from this analysis. With no prospect of

material cost declines, small-scale wind is not considered further in this report.

Figure 18 – Grid-Scale Solar LCOE Forecast to 2030

Figure 19 – Wind Capital Cost Forecast to 2040


28


LCOEs are available from Lazard, EIA and NREL. A long-term LCOE forecast for grid-scale wind was only

available from the EIA and NREL. NREL is forecasting 1.6% to 1.8% LCOE declines for grid-scale wind. The

EIA is forecasting a 10% cost increase. To be conservative, NREL’s aggressive cost declines have been

assumed for this analysis. NREL predicts a future LCOE of $37/MWh in 2030, down from $46/MWh in

2017, for a wind farm with a capacity factor of 41% today. Lazard’s estimate for a wind farm today is

similar to NREL’s but assumes a much higher generator capacity factor of 47%. This places the Lazard

estimate between the EIA and NREL estimates of common capacity factor assumptions. By choosing the

NREL reference point, this report may be understating the future cost of wind by 15%.

4.2 Storage Costs

This section examines the forecasts for the capital cost and LCOE of storage technologies from multiple

sources. For each technology, different deployment scenarios scales are considered from small-scale

residential applications to large grid-scale. Three distinct storage technologies are addressed:

• Li-ion batteries

• Pumped hydro storage

• CAES

The cost comparisons for both capital and LCOS are provided in Figures 20 and 21.

Capital costs for storage are measured in two ways: (1) In $/kW, which refers to the current carrying

capacity of the input/output electronics required to charge or discharge energy (similar to how a

generation plant is measured); and, (2) In $/kWh, which measures the volume of energy that can be

stored. In this report, capital costs are compared on a $/kWh basis. Li-ion batteries are more expensive

today than pumped hydro or CAES, but with the significant projected capital cost declines may begin to

approach the capital cost for pumped hydro.

Figure 20 – Storage U.S. Average Capital Cost, 2017 vs. 2030


29


The primary measure used in this study for assessing the cost of storage options is the LCOS49. Based on

this measure, Li-ion batteries are expected to remain more expensive than pumped hydro or CAES. Small-

scale residential storage is expected to remain above $400/MWh suggesting that residential based DER

solutions will not be economic for well beyond the time horizon of this analysis.

4.2.1 Lithium-Ion Storage Costs

Lazard50 is the primary source consulted for Li-ion storage costs and their expected cost declines over the

next five years. Three scales of storage have been identified:

• Residential (5 kW)

• Microgrid (1 MW), which is considered as “community” in this study

• Distribution (10 MW)

a) Capital Cost Forecasts

The capital cost for an energy storage system (ESS) is comprised of the storage module (SM), balance of

system (BOS), power conversion system (PCS) and related engineering, procurement and construction

(EPC) costs as illustrated in Figure 22.

49 The LCOEs shown for all storage devices are assumed to provide 8 hours of storage and to be fully charged and discharged every day for 350 days per year. 50 Lazard LCOS v3.0

Figure 21 – Storage U.S. Average LCOS, 2017 vs. 2030


30


The contribution of a cost category to LCOS varies by use case and technology.

Lazard summarizes the three capital cost components of a Li-ion Battery system:

• Initial capital cost – Direct Current (DC)

• Initial capital cost – Alternating Current (AC)

• Initial other owner costs

Table 4, based on Lazard’s data, summarizes the respective share of each of these cost components.

Table 4 – Lithium-Ion Battery Capital Cost Components

Lazard forecasts that the capital costs of the storage technologies will decline by 10%/year for the next

five-years. Lazard also notes that the BOS costs would decline more in line with those of solar. The proxy

for solar cost declines was obtained from the Leidos report for large-scale commercial installations.

Applying the 10% per year Lazard cost decline to the DC and AC components for the next 5 years and using

solar CAGRs for the rest to 2030, yields CAGRs for the total installed cost of approximately 8% to 9% per

year51 across all technology scales as summarized in Figure 23.

51 Recent forecast released by GTM suggests storage cost declines will be 8%/year for the next 5 years suggesting that the costs projected here may be aggressively lower than others may expect. IRENA 2017 projects that 2030 costs will be 60% lower than 2016, also not as aggressive as the forecast developed in this report.

Distribution (10 MW)

Microgrid (1 MW)

Residential (5 kW)

Li-ion Battery (dc) 347 480 646

Li-ion Battery (ac) 19 39 314

Other owner costs 55 104 200

Total 421 623 1160

Table 4 - Lithium-Ion Battery Capital Cost Components

($U.S./kWh, 2017)

Figure 22 – Lithium-Ion Battery Physical Energy Storage System


31


The resulting capital cost forecast for the three scales of storage is provided in Figure 24.

b) Levelized Cost of Storage

Lazard identifies several cost components that contribute to the LCOS of Li-ion battery storage. These

components are illustrated in Figure 2552, and include the following four categories:

• Capital costs consisting of the AC, DC and EPC components

• Augmentation costs, which account for additional investments that will be required to sustain

performance as the battery life degrades

52 Copied from Lazard LCOS V3.0

Figure 23 – Lithium-Ion Battery Capital Cost CAGR Forecast to 2030

Figure 24 – Lithium-Ion Battery Capital Cost Forecast to 2030


32


• Operation costs, which include operations and maintenance (O&M), warranty, and charging

• Other costs, e.g. debt service (or capital financing costs) and taxes

For the LCOS used in this report, taxes and charging costs are excluded.

Previous Lazard forecasts did not include augmentation costs which are now known to be material,

particularly for small installations. Augmentation costs represent the additional ESS equipment needed to

maintain the “Usable Energy” capability: cycling the unit according to the usage profile for the life of the

system. Useable Energy may degrade under the following circumstances:

• Unit is not charged/discharged at the planned rate (kWh) per cycle

• Battery chemistry does not have the cycle-life needed to support the desired operating life

• The energy rating (kWh) of the battery chemistry degrades over its life

In assessing the cost of an ESS upgrade, Lazard has taken into account the falling price of ESS system costs.

The LCOS forecast was developed as follows:

• For the capital portion of LCOS, the same CAGRs are used as for the capital cost. As augmentation

and debt service costs are bot related to capital, it is assumed that they will also see similar

declines. This is likely an aggressive assumption for the augmentation costs as they have already

been forecast out to the future where ingoing declines will be more modest.

• For the operating cost of the LCOS, CAGRs from the Leidos fixed O&M for large commercial solar

PV have been applied.

The resulting LCOS CAGRs range from 6.6-8.3% for the period of 2017-2020 and drop to 3.8-4.5% for 2025-

2030 as shown in Figure 26. CAGRs are higher for residential applications due to the predominance of

capital costs.

Figure 25 – Illustrative System Costs: LCOS by Category ($U.S./kW-yr)


33


The net LCOS forecast is shown in Figure 27, reflecting a cost of $111/MWh for the community- scale

(microgrid) Li-ion battery solution in 2030.

4.2.2 Pumped Hydro and Compressed Air Energy Storage Costs

Pumped hydro and CAES technologies have existed for several decades. The capital costs and LCOS of

pumped-hydro and CAES were obtained from Lazard which suggests that capital costs will decline by less

than 1%/year. To compute CAGRs for the declining costs of these storage options, the EIA and NREL

forecasts for hydro resources were used for pumped hydro and a blend of CAGRs from conventional

sources was applied to CAES.

Figure 26 – Lithium-Ion Battery LCOS CAGR Forecast to 2030

Figure 27 – Lithium-Ion Battery LCOS Forecast to 2030


34


Figure 28 shows modest declines between 2017 and 2030 of 0.6% and 0.4% per year in average forecast

costs for pumped hydro and CAES respectively.

4.3 Integrating Renewables with Storage

In order for DER solutions to provide the supply required to meet the expected demand profile, the

technologies must be integrated. As discussed earlier, two demand profiles must be satisfied: baseload

demand; and, daytime demand.

Integrating storage and generation presents both cost and performance implications that are determined

by several factors:

1. Configuring the components (e.g. the size of the solar panel and capacity of the battery) to supply

the expected demand profile determines the net blend of the directly used electricity or electricity

obtained from storage produced by the DER option.

2. Energy losses arising from the round-trip conversions from charging and discharging arise within

the storage system.

3. Integrating solar and storage systems through coupling of the electrical components could reduce

the combined capital cost.

These parameters yield the LCOE of the system outputs as illustrated in Figures 29 and 30. First, the raw

renewable generation cost estimates must be adjusted to account for the connection costs to the grid

and to identify the cost of the generation technology employed. The generation that gets stored has a

cost increase due to the losses in storage and would also bear the cost of the storage itself. Ultimately,

the net cost of energy from the integrated DER system is the weighted average of the directly used and

stored energy.

Figure 28 – Pumped Hydro and Compressed Air LCOS Forecast to 2030


35


The results illustrated above reflect a usage cycle that sees the storage fully charged and discharged on a

daily basis. As such, even without consideration of the intermittency of the renewables, it can be expected

that the cost of DER in the future will be double the cost of the renewable inputs.

4.3.1 Solar-based DER Solutions

Figure 31 illustrates the breakdown of cost impacts of solar-based DER required to meet a community-

scale daytime demand profile. Four steps include: (1) assessing the cost of the utilized solar output; (2)

the costs of the stored solar; (3) the cost/benefits of the storage from the integrated components; (4) and

value of the net blended outputs.

Figure 30 – Integrated Wind + CAES LCOE

Figure 31 – Solar Community DER LCOE Components

Figure 29 – Integrated Solar Community DER LCOE


36


a) The Cost of the Utilized Solar Output

When solar generation is installed, there are grid connection charges associated with the installation. The

EIA adds two cost factors to their LCOEs when forecasting future costs. The first is a technology uncertainty

factor and the second is the cost of Tx. For large scale renewables, their low capacity factor results in

unused Tx, especially for the radial connections from the wind and solar farms to the load centres. That

all adds additional delivery costs to intermittent renewable supply options. For solar, the EIA factors add

~$8/MWh to the LCOE which have been assumed to be relevant for this study.

Solar grid integration costs are relatively high due to the expected intermittency of the solar output while

the system must be able to accommodate 100% of the peak output. For integrated DER solutions, it is

assumed that the peaks will be absorbed by storage and hence the grid costs would be reduced. For the

purpose of this analysis, 50% of the grid costs have been removed, reducing the forecast cost of

community-scale solar by $2.2/MWh. The cost of solar output that is utilized is expected to be $77/MWh.

b) The Cost of Stored Solar

The cost of stored solar is a function of the losses that occur when the solar energy is used to charge the

storage and when the storage device discharges the energy back to the system. This round-trip efficiency

loss for Li-ion batteries is estimated at 14%. With a utilized solar energy cost of $77/MWh, the cost of the

stored solar energy will be 14% higher, or $11/MWh more. This results in a combined cost of $88/MWh,

before including the cost of the storage technology itself.

c) Cost/Benefits of Integrated Storage with Solar Panels

NREL conducted a study53 on the possible benefits of integrating the electronics of the solar panel with

that of the storage device. The study found that coupling solar PV and storage can reduce costs by sharing

components in two ways:

• AC-coupled system shares little hardware, but costs associated with engineering, customer and

site acquisition, permitting, and labour can be shared; and,

• For a DC-coupled system, the second inverter can be eliminated to reduce costs

In a tightly integrated arrangement where the storage device is only charged with energy from the solar

panel, the cost of the BOS systems can be reduced.

The conceptual arrangement of a DC coupled system that permits some flexible charging is illustrated in

Figure 32.

53 NREL, Evaluating the Technical and Economic Performance of PV Plus Storage Power Plants, 2017


37


The cost/benefits summarized in Figure 33 were extracted from the NREL report. This data suggests that

40% of the battery BOS capital costs could be avoided.

Estimates have been developed for how this integration may impact the different sizes of possible solar

installations as summarized in Figures 34 and 35.

Applying the savings to the LCOS in Figure 35 will reduce the per unit cost of storage for a community

solution by $5/MWh from $111/MWh to $107/MWh.

Therefore, the total cost of solar energy from storage in a community DER system is expected to be

$194/MWh in 2030.

Figure 32 – Conceptual Arrangement of a DC Coupled PV + Storage Battery System

Figure 33 – Avoided Cost of Battery BOS Associated with PV Coupling ($/kW)

Figure 34 – Lithium-Ion Battery Capital Cost Reduction

Figure 35 – Lithium-Ion Battery Integration LCOS Reduction


38


d) Determining the Net Blended Cost

To determine the net blended cost of energy from a DER system requires a model of system usage. To

effectively pair with renewables, an optimum storage function would create an output that matches the

demand profile. Two demand profile scenarios are illustrated in Figures 36 and 37: daytime demand; and,

baseload demand.

In the daytime demand supply scenario, storage and solar combine to emulate a gas plant supply

alternative. For an average Ontario day in September, storage would be sized to capture 52% of solar

energy, with 45% used to charge the battery after 7% losses. September is a useful month for this analysis

as it has 12 hours of sunlight per day. The net cost of blending the 48% of solar that is directly used with

the 45% from storage is $134/MWh. This is based on the assumption that the storage system is fully

charged and discharged on a daily basis.

To supply the baseload demand scenario, storage and solar combine to provide a 24x7 supply, emulating

a nuclear plant in Ontario. For the same average Ontario day in September (12 hours of daylight), storage

would have to be sized to capture over 60% of the solar energy.

To provide a baseload supply solution, 20% more storage would be required. This is because the solar

output needs to be spread out at a lower level over a much longer period of time along. This results in an

increase in the system LCOE to a cost of over $150/MWh. For this reason, solar baseload DER solutions

are not considered further in this analysis.

4.3.2 Wind-based DER Solution

Figure 38 illustrates the breakdown of cost impacts of wind-based DER required to meet a community-

scale daytime demand profile. This is based upon: (1) an assessment of the cost of the utilized wind

output; (2) determination of the costs of the stored wind; (3) computation of the net blended outputs.

Figure 36 – Solar and Storage Daytime Demand Supply

Figure 37 – Solar and Storage for Baseload Supply


39


a. The Cost of Utilized Wind Output

As with solar generation, there are grid connection charges associated with wind installations. The low

capacity factor for large-scale renewables results in unused Tx especially for the radial connections from

the wind and solar farms to the load centres. In the case of wind, the length of the radial lines is typically

longer than that for gas or nuclear generation. This adds additional delivery costs to intermittent

renewable supply options. For wind, the EIA Tx and contingency factors add $4/MWh to the LCOE.

On a LCOE basis, wind grid integration costs are relatively lower than for solar due to its higher capacity

factor. For integrated DER solutions, it is assumed that the peaks will be partially absorbed by storage and

hence the grid connection costs will be reduced. For the purpose of this analysis, 50% of the grid costs

have been removed, reducing the forecast cost of used wind by $1.4/MWh. The cost of utilized wind

output is expected to be $40/MWh.

b. The cost of stored wind

The cost of stored wind is a function of the losses that occur when the wind energy is used to charge the

storage facility and then when the storage device discharges the energy. This charge/discharge cycle

results in an efficiency loss estimated to be 35% for CAES. For utilized wind output that costs $40/MWh,

the stored wind energy cost will be 35% higher, or an additional $14/MWh. This yields a combined cost of

$54/MWh before the cost of the storage itself is added. The LCOS of CAES, assuming a full duty cycle, is

projected to be $63/MWh. This suggests that the cost of the energy retrieved from storage to be

$117/MWh.

Figure 38 – Wind + CAES LCOE Components


40


c. Determining the Net Blended Cost

Figures 39 and 40 illustrate the design parameters for wind-based DER for the two demand profiles.

In the daytime demand supply scenario, storage and wind are combined to emulate a gas plant supply

alternative. For the average Ontario day in March, storage would be sized to capture 20% of wind energy.

March is characterized as a relatively high demand month and also a high wind production month. A net

blended cost of $57/MWh is expected based on a 70% utilization of wind output and 20% consumption

of stored wind. This assumes that the storage system is fully charged and discharged on a daily basis.

In the baseload demand supply scenario, storage and wind combine to provide a 24x7 supply, emulating

a nuclear plant in Ontario. For the same average Ontario day in March, storage would have to be sized to

capture only 4% of the wind energy.

To provide a baseload supply solution, much less storage may be required and could result in a lower

system LCOE. For this reason, wind-based DER baseload solutions will continue to be considered in this

analysis.

4.4 Case Examples of DER Economics

Examples in this section illustrate how DER is being justified today as viable for consumers who install

them. Consumer benefits are typically achieved through pricing mechanisms that inherently push costs

onto other ratepayers. DER case examples show that the underpinning premise of DER adoption is the

economic benefit that can accrue to the owners of DER systems. Those owners take advantage of two

benefits: peak pricing (arbitrage) in a fossil-fuelled energy system; or from indirect subsidies (e.g. net

metering). Both of these opportunities benefit DER owners but add costs to the system that are borne by

other ratepayers.

This study articulates renewables-based DER economics from the following perspectives:

1. Providing advantages to the system

2. Leveraging market pricing to maximize arbitrage

3. Maximizing renewables penetration

Figure 39 – Wind Capacity Sizing for Daytime Community Demand

Figure 40 – Wind Capacity Sizing for Baseload Demand


41


4. Combining technologies to provide system solutions

4.4.1 Advantages to the System

Much debate surrounds how to define and capture the value and monetize the benefits of DER for the

system. Developers also expect an acceptable rate of return for their DER investments. The state of

Massachusetts has demonstrated that most of the DER system benefits cannot be captured by the end

user through normal market mechanisms54. The Essex report previously discussed in Section 3.0 also

identified that consumers can only capture some of the benefits.

NREL’s report on DER compensation models55 stresses the importance of developing well-designed

mechanisms to help minimize the negative impacts and maximize the value of DER to all stakeholder

groups, including Dx utilities, the DER system owner, and other ratepayers (non-DER-system owners). Like

with the Massachusetts report, NREL indicated that it is a relatively straightforward exercise to identify

the direct benefits to the DER system developer. NREL also recognized that it is far more difficult to

quantify the longer-term benefits to the system in a manner that can be built into a compensation model

for the DER developer. If the net effect of DER adoption is determined to be a net cost, then non-adopting

customers can be expected to see an increase in their bills. One of the compensation models NREL defined

is net metering.

Finding a resolution to the compensation question is fundamental to resolving this debate and fairly

allocating the costs/benefits to the system, DER developer/installer and non-DER consumers. In California,

the widespread prevalence of rooftop solar panels with net metering has shifted costs from households

with residential solar panels to those without. The cost is estimated to be $65 per year for the average

household56. As noted earlier in this report, distributed solar proponents suggest DER allows utilities to

benefit by avoiding or deferring Dx system upgrades, yet these impacts have been found to be relatively

small. The New York Public Commission is developing new methodologies to value and compensate DERs

that take into account energy price, cost reductions for consumers, and the value of deferred

capital57. Regulators in California are exploring models, such as locational net benefits, to facilitate the

transition to a distributed energy future58.

4.4.2 Leveraging Market Pricing to Maximize Arbitrage

To date, most DER systems have been implemented either under PPAs (e.g. Tucson Electric and NextEra

in Arizona59) or are based on leveraging the market pricing. The latter is typically driven by electricity

systems in jurisdictions were fossil-fueled generation predominates. The desired outcome of leveraging

market pricing is to maximize revenues, leaving it to the grid to fill in the low-cost gaps. In this report, the

54 Massachusetts DOER, State of Charge, 2016 55 NREL, Grid-Connected Distributed Generation: Compensation Mechanism Basics, 2017 56 Davis, 2018 57 State of New York. Public Service Commission. Order on Net Metering Transition, Phase One of Value of Distributed Energy Resources, and Related Matters. 2017 58 California. Public Utilities Commission. California's Distributed Energy Resources Action Plan: Aligning Vision and Action. 2016. Pages 1-14 59 Maloney, 2017


42


design of DER systems to supply a demand profile does not assume that “grid” supplies will be available

to provide that function.

Three models illustrate how value is captured by DER installations through maximizing revenues from

market prices versus the delivery of energy to meet a demand profile.

a) Net Metering

Ontario’s LTEP describes the virtues of Net Metering and provides an example of how net metering can

be used to generate cost savings for DER owners. The LTEP is silent about the cost implications for the

total system. The LTEP defines net metering as a billing arrangement that allows consumers to generate

their own electricity on site for their consumption and also receive energy credits for any extra electricity

that is delivered to the local Dx system. This arrangement is illustrated in Figure 4160.

The energy credits are then subtracted from any energy the customer may later draw back from the grid.

This approach is more fully illustrated in Figure 4261.

60 MoE LTEP, 2017 61 NREL Grid-Connected Distributed Generation, 2017

Figure 41 – Residential Net-Metering with 4 kW Rooftop Solar PV


43


In this manner, a PV owner in Ontario is avoiding paying the global adjustment. Other benefits that depend

on jurisdiction include volume dependent Dx, Tx and regulatory charges for the credited energy receive

back from the grid. This bill avoidance is an indirect subsidy that other ratepayers will be providing if there

are no benefits to the grid from the net-metered DER.

There are several challenges with the net metering approach for solar installations.

• The excess solar will be put on the grid when electricity prices are lower and will be drawn back

when electricity prices are higher. That is because, as illustrated in Figure 41, the solar peak is not

coincident with system peak. Ontario is already suffering from many instances of surplus solar

energy62 which drives down further the price commanded mid-day.

• The expected benefits to the system of avoiding or deferring Dx system investments are unlikely

to be realized as peak demand is not affected by solar output since it is not coincident. If the DER

system is unable to offset peak demand, then system benefits will not accrue.

• The payers of the indirect subsidy are the ratepayers who do not have solar-based DER.

The introduction of rooftop solar is a net extra cost burden on the electricity system and will continue to

drive up energy costs through both subsidies and system issues created from its intermittency as

described by the IESO.63 The OEB has recognized that net metering will not be offsetting Dx system costs

and has directed that delivery and regulatory costs will be converted to a fixed charge for LDCs in Ontario.

This will be implemented over a four-year period. Other ratepayers would still have to absorb the

unrecovered global adjustment costs.

The LTEP is also supporting the additional notion of virtual net metering to allow Ontarians who may not

be able to install their own renewable energy system to participate in renewable energy projects located

away from their homes or businesses, and still receive a credit offsetting their electricity bill. Unless

storage solutions can economically ensure peak shaving, there can be no system benefits. Unfortunately,

the use of storage in a net-metering context is in conflict with the benefits of net metering to solar panel

62 IESO data shows that 20% of the solar generation output from the recently installed grid solar farms was curtailed in 2017 63 IESO, Energy Storage, 2016

Figure 42 – Net Energy Metering


44


owners because it would limit the price arbitrage benefits. Net metering to support the addition of high

cost generation in Ontario is a bad approach.

b) NREL Models for Solar PV Plus Storage Systems

NREL undertook a study to assess the benefits of integrating solar PV with storage64. They considered

several different architectures and tested those architectures against revenue models related to the

energy markets in California. All of their revenue models involved attempting to capture the highest value

of energy offsets from the PV plus storage system. The purpose of illustrating these cases in this report is

to emphasize that the design considerations are not related to demand profiles being served, but to

maximizing the revenue potential in a market dominated by fossil fuel prices.

Case 1 – Tightly Coupled PV + Storage System

A tightly coupled system is one where the PV and storage systems use a common set of electronics and

the batteries are only charged with energy from the solar PV panels. Figure 43 illustrates the charging and

discharging strategy for a tightly coupled PV + storage system in California.

The chart on the right shows how the solar output is used to charge the batteries in the morning and then

discharge them in the afternoon. The chart on the left shows how the storage is being charged during

periods of low system marginal prices and discharged during times of high system marginal price, hence

maximizing the revenue potential for the system owner.

Case 2 – Independent Storage and PV systems

Independent systems do not have integrated shared circuitry. These systems can be individually optimized

against market signals. Figure 44 shows how the storage is charged during low cost overnight periods and

then discharged during high price periods. The solar is just used when it is available.

64 NREL, Evaluating the Technical and Economic Performance of PV Plus Storage Power Plants, 2017

Figure 43 – Optimal Dispatch of a DC Tightly Coupled PV plus Storage System (June 16, 2014 Price Data from SCE LAP)


45


In this case, the storage is not being used to support the solar output in any way, it is being used to

independently maximize the revenues of the storage based on market prices.

c) Valuing Energy as a Function of When it is in Demand

Early studies examined the value of renewables output in the context of the market value of energy at the

time the renewables outputs were available65. The argument presented is that the industry standard LCOE

measure for comparing energy sources should not be blindly applied to energy sources such as

renewables because their output does not necessarily align with demand. This is a particularly relevant

perspective when considering that wind output occurs at any time of the day. Solar output occurs during

the middle of the day and so would normally not be impacted by very low market prices; however, surplus

solar conditions are emerging in Ontario requiring the curtailment of solar resources. To identify the

“value” of this energy source, the market value of energy at the time the renewables are generated could

be a criterion for defining the LCOE for comparison purposes.

Renewables’ intermittency leads to situations where that energy is simply surplus. The wasted surplus

costs are not reflected in this LCOE treatment as the system in which the renewables are inserted impacts

on that value. As low-emission generation is pursued, the market price, which is based on the variable

cost of generation, is becoming less relevant and not useful as a trading mechanism. This is creating much

discourse in the U.S. on how markets require restructuring to address the implications on the future of

electricity supply in that country.

4.4.3 Models to Support a Vision for Maximizing Renewables Penetration

The integration of the many components of an energy system have been looked at to estimate how that

system may respond to renewables’ intermittency and still deliver the energy demanded by the

jurisdiction. These approaches look at the system as a whole and consider optimization benefits that could

reduce overall costs.

65 Joskow, 2011

Figure 44 – Storage Charging During Low Cost Overnight Periods


46


a) Maximizing Renewables in an Energy System

Models of how renewables can form the basis for an entire energy system have been undertaken66 with

the purpose of assessing the total system cost and not considering pricing implications. The underlying

premise for the analysis is the perspective that renewables “should be the solution” and that other

resources (storage and natural gas) would be drawn upon to accommodate the shifting of supply to meet

demand when the intermittency results in misalignment. The design of such a system for Germany

entailed coupling wind and solar resources sized to deliver the full energy demanded over a year with a

mix that minimized the seasonal storage needed. Battery storage was based on the average daily energy

shifting required to match the renewables to demand. Modelling of generation and demand profiles

indicated that 14% of the renewable energy had to be curtailed, 6% of the renewables were shifted by

storage (with the associated losses), and natural gas fired generation was relied upon to supply 14% of

the demand. Most of the natural gas use was required to address the seasonality implications. Of note is

that natural gas fired generation only represented 4% of the supply mix in Ontario in 2017, a year in which

24% of the renewables supply was curtailed.67

Notwithstanding the wasted renewable energy and need for natural gas, the findings suggested that this

mostly renewables system could supply Germany’s energy needs at less than the cost of a system that

relied solely on natural gas, even with a $50/tonne carbon price. However, the German study used: (1)

overly optimistic costing assumptions; and (2) simplified generation and demand profiles which did not

adequately reflect the impact of intermittency, the most significant factor affecting cost.

b) Advanced data analytics for system optimization

The science of advanced data analytics, and indeed artificial intelligence, is impacting the electricity

sector. The concept of microgrids being optimized through balancing demand and supply for the

community is yet another innovation in the toolbox for utilities. Smart controllers coupled with smart

storage offer the potential to optimize energy resources across a community and marry renewables

intermittency to demand fluctuations. It has been postulated that data analytics could also enable

effective sharing of energy resources across the grid.

In terms of meeting the demand of a community when the local microgrid has insufficient self-generated

energy, relying on sources outside the community does not meet the requirements for DER set out for

this analysis. Furthermore, neighboring communities in Ontario have weather patterns that are very

similar and hence unlikely to create much of a difference between them in surplus, stored, and need for

backup energy.

4.4.4 System Solutions Must Combine Technologies

For systems to optimally match energy supply to demand, multiple technologies need to be engaged. New

ICT-based smart control technologies are facilitating the integration of renewables and storage

technologies enabling two new paradigms: (1) a new class of customers: consumers and producers of

electricity (“prosumers”); and (2) community-based microgrids and virtual power plants. DER connected

66 CPI, 2017 67 IESO Year End Review 2017


47


in a microgrid configuration has the potential to provide at the local Dx system level the dispatch flexibility

that natural gas-fired generation currently provides for the Tx grid.

Storage may be the enabler for adapting the energy supply to the demand, as it can intermediate between

the supply and energy demanded by users. The advanced controllers and data analytics discussed earlier

can also help optimize:

1. Demand fluctuations by encouraging users to adjust their usage profile (e.g. automation that

enables EV charging or water heating at night, or curtailment of daytime air-conditioning when

other energy usage such as for refrigerators and freezers is high; or managing options for hybrid

electrical/gas appliances) to minimize impacts on system peaks.

2. Using storage to follow peaks in demand regardless of how supply is used to charge the batteries

(as long as there is enough supply).

The impact of advanced controllers and data analytics may lead to flattening or smoothing of the daily

demand load minimizing the impact of demand fluctuations on energy supply. Flattening the daily demand

profile may also attenuate the drivers that cause higher prices in the market.

4.4.5 Summary Observations

Methods for establishing the business cases for DER that are predicated upon market pricing practices

generally favor installers of DER at the expense of increasing total system costs. These case examples

assume the fossil system will fill in any gaps required to meet demand, an approach which ignores the

need to account for the cost of backup generation to complement renewable solutions.

Considering the LCOE of the individual components is insufficient for gaining a fair comparison. Just

because grid-scale solar is becoming less expensive does not directly imply that an integrated grid-scale

solar system will be cheaper. Examining the full system costs of introducing DER allows for a proper

assessment of the impacts to the overall system. However, assumptions and modeling practices should

be carefully selected to properly illustrate the implications.

The advent of advanced data analytics, controllers and artificial intelligence (AI) have the potential to alter

the demand profiles placed on DER systems by flattening the demand and reducing peaks. Such

innovations will tend to favor energy supplies that have reliably flatter or more baseload type

characteristics.

4.5 Comparison with Conventional Generation

Understanding the cost of renewable-based DER systems is most informative when those costs can be

compared to alternatives. This section summarizes the costs of conventional solutions to the baseload

and daytime demand challenges as well as a nuclear baseload-supplied DES. Consideration is given

primarily to low emission options which include nuclear and natural gas fired generation equipped with

carbon capture.

Figure 45 summarizes the forecast costs for these conventional solutions. Nuclear technologies would

provide a baseload supply but can also serve daytime demand if coupled with distributed storage. For


48


supplying daytime demand, CCGT equipped with CCS may be competitive with other forms of gas-fired

generation if carbon pricing is included (carbon price of CAD $100/tonne illustrated).

This section addresses three topics:

1. The forecast costs of conventional nuclear and CCGT with CCS

2. The cost implications of a nuclear baseload-supplied DES option

3. The costs of alternative gas-fired generation options.

4.5.1 Cost Forecast for Conventional Generation

For conventional generation, the EIA and NREL forecast LCOEs were considered as shown in Figure 46 for

nuclear and Figure 47 for CCGT equipped with CCS.

Figure 45 – LCOE of Nuclear and Natural Gas-Fired Generation


49


Both capital costs and LCOEs for these technologies are expected to have modest declines of much less

than 1% per year. While the EIA and NREL have current estimates for CCS, these estimates grow in the

future before starting their modest declines. For that reason, only the 2030 values are shown. The small

modular reactor (SMR) estimate is based on the Energy Information Reform Project (EIRP)68.

68 EIRP, 2017

Figure 46 – Nuclear LCOE Forecast to 2030

Figure 47 – Combined Cycle Gas Turbine with CCS LCOE Forecast to 2030


50


4.5.2 Nuclear Baseload-Supplied DES Option

The full cost impact breakdown for nuclear coupled with storage to meet the community-scale daytime

demand profile is illustrated in Figure 48. These include: an assessment of the cost of the nuclear energy

utilized; the costs of the nuclear output that is stored; and a computation of the net blended outputs.

a) The Cost of Utilized Nuclear Electricity

As the nuclear baseload-supplied DES option provides baseload supply to the grid, the cost of the utilized

nuclear electricity includes the full EIA markup for Tx and contingency.

b) The Cost of Stored Nuclear Electricity

The storage for nuclear baseload-supplied DES is assumed to be Li-ion batteries. The full cycle round-trip

efficiency loss is 14%. For the utilized nuclear electricity at a cost of $94/MWh, the cost of the stored

nuclear electricity will be 14% higher, or $13/MWh more, yielding a total of $107/MWh before the cost

of the storage component is added. The LCOS of the batteries is $111/MWh, which is higher than for solar-

based DER which benefitted from technology integration economies. The net cost of the stored nuclear

electricity would be $218/MWh.

c) Determining the Net Blended Cost

The design parameters for a nuclear baseload-supplied DES solution required to supply the daytime

demand profile are shown in Figure 49. The illustration reflects a community DES with 1.2 MWh of storage.

Figure 48 – Nuclear Baseload-Supplied DES LCOE Components


51


To supply the daytime demand scenario, storage and nuclear are combined to emulate a gas plant supply.

For the average Ontario day in September, storage would be sized to capture 23% of the nuclear energy.

September was chosen as a sample month to facilitate a comparison with the solar option. The net cost

of blending the 77% of nuclear output that is directly utilized plus 20% output from storage totals

$120/MWh. This assumes that the storage system is fully charged and discharged on a daily basis.

4.5.3 Natural Gas-Fired Generation Options

While CCGT with CCS has been chosen as the low-emission grid-based supply option for contrasting cost

performance of DER solutions, other natural gas-fired options may also be desirable depending on the

climate policies in place. It may be desirable for cost reasons to accept a high emitting generation option

that has a low operating factor and simply pay the carbon price.

In the DER discourse, one of the options considered is microturbines that can be located within community

or distribution-scale DER installations.

Figure 50 summarizes the LCOE for several forms of natural gas-fired generation. The left side of the figure

shows the LCOE of the gas-fired generation options at the nominal capacity factors assumed for the prices

quoted. These are typically the prices referred to when these natural gas-fired generation options are

compared. To choose among them for the purpose of this study, the comparison is best based on common

capacity factors.

Figure 49 – Nuclear and DES for Daytime Supply


52


The right side of the figure compares the LCOE of each generation type for an installation with a 30%

capacity factor. Carbon pricing has been included for two levels: CAD $50/tonne; and CAD $100/tonne.

Canada is expected to have a $50/tonne carbon pricing mechanism in place by 2022. Carbon pricing will

need to be well in excess of CAD $100/tonne by 2030 if emission reduction objectives continue to be

pursued69. With carbon pricing considered, the large-scale gas options have similar total cost implications,

but clearly have different emissions profiles. CCGT with CCS could cost 15% more than regular CCGT but

would produce 90% less emissions.

The capital portions of the LCOEs are an indicator of the sensitivity to the capacity factor. The capacity

factor indicates how much the gas plant is used. The small capital component for peaking generators

makes them the least expensive option at low capacity factors. The high capital content of the CCGT with

CCS option suggests that their competitiveness will improve as capacity factors increase.

The microturbines stand out as the most expensive backup supply option even without carbon pricing

given their high capital requirements and high variable fuel consumption. Microturbine solutions for DER

applications are likely cost prohibitive from a total system cost perspective.

4.6 Summary of Cost Implications

The research of DER component costs suggests that costs will continue to decline, although not at a rate

of acceleration that will make renewables and storage ubiquitously cheap. By 2030, grid-scale wind is

expected to decline by 20%, community-scale solar could decline by over 35%, and storage by about 50%.

69 Strapolec, Emissions and the LTEP Phase 2, 2016

Figure 50 – Natural Gas-Fired Generation LCOE Comparison


53


When these technologies are integrated into a DER solution, the total costs will be much higher, even

under ideal conditions. Today, costs are far too high for DER solutions to be considered on a large scale

for managing system peak demand. By 2030, it appears that grid-based wind solutions may be economic

when compared to nuclear or natural gas solutions. The economics of community based solar DER,

however, may be questionable even though some capital cost synergies could be realized.

The current high cost of DER calls into question how DER solutions are being adopted in the market place.

Most solutions are predicated on market arbitrage opportunities and rarely consider total system costs.

There is much literature addressing the challenges of properly valuing and monetizing the costs of DER.


54


5.0 Reality of Intermittency

Renewables intermittency is a function of how sunny it is and how much the wind blows.

This intermittency undermines the efficiency of a DER solution, increasing the cost of both useable

renewable energy and storage. In a solar-based DER system, the solar panel and storage capacity must be

physically sized to a set of assumptions, which would typically reflect an average output expectation. The

production variability associated with renewables impacts the efficiency and costs of the generation and

storage components. These impacts are manifested in several ways:

1. When renewables output exceeds storage capacity, this creates surplus output and the need for

curtailment.

2. When renewables output falls below the expected capacity factor, two consequences can occur:

a. Gas backup is required to meet the demand.

b. Insufficient energy is available to fully charge the storage device, causing the storage capacity

factor to drop.

Demand fluctuations can also impact the efficiency of the DER system. Similar to renewables’

intermittency, low or high demand can impact the need to curtail the output from generation, the need

for backup supply, and the use of storage capacity.

Figure 51 illustrates the effects of these two factors on a DER system. The characteristics of a system based

on the average output that assumes no intermittency, as described in Section 4.4, is contrasted against

the impacts of the intermittent nature of renewables and the reality that demand also fluctuates. Wasted

generation refers to both losses in the storage system as well as any surplus generation that would get

curtailed. As such, even the perfect no-intermittency case has wasted energy in the form of storage losses.

Section 5.1 discusses the various factors that result in intermittency and demand fluctuations and

presents the demand requirements that DER systems in Ontario would have to meet.

Figure 51 – Intermittency and Demand Fluctuation Impact on DER Component Use


55


Section 5.2 discusses the causes of solar intermittency while Section 5.3 looks at wind intermittency.

Section 5.4 characterizes the impacts of demand fluctuations on solar-based and wind-based DER, as well

as on nuclear baseload-supplied DES, and compares the results.

5.1 Intermittency Defined and Characterized with Demand

It is well understood that the generation profiles of solar and wind output do not match the electricity

demand profile, as illustrated by Figure 52.

Solar peaks in the spring when demand is at its lowest. Wind generation is highest in the winter, matching

demand, but then is too high in the spring and fall, and too low in the summer when demand peaks again.

Storage is considered to be a solution for addressing this mismatch between output and demand and to

facilitate the integration of renewables. The storage system would charge during hours of excess

renewable generation and discharge during periods of low or no generation. This concept was illustrated

for wind earlier in Figure 39, without consideration of the impacts of intermittency.

In addition to the misalignment with demand, wind and solar generation present a significant

intermittency challenge – significant output being generated one day and none on another. Demand

variability on a daily and seasonal basis represents another challenge.

To properly compare the effects of both supply intermittency and demand fluctuations on DER options,

these intermittency and demand fluctuations need to be characterized and a common set of requirements

defined that each DER option would be expected to meet.

This section describes the types of intermittency and the nature of demand fluctuations and establishes

the requirements for future DER systems in supplying the expected demand for Ontario.

Figure 52 – Relationship Between Wind, Solar, & Community Demand


56


5.1.1 Renewables Intermittency and Relationship to Storage

Intermittent outputs from wind, solar and hydro generation impact the value of these renewable assets

and also the potential value of the storage used to smooth their outputs. Storage capacity utilization is

“cost optimized” when the storage is charged and discharged more frequently. Most storage costing

assumptions assume a daily charge/discharge cycle for most days of the year. Table 5 summarizes the

relationship between storage and renewables intermittency by type.

Table 5 – Types of Renewables Intermittency and Variable Generation

Type of Intermittency Impact on Storage

Hour to hour Can be smoothed by storage, it is the easiest job storage can do and aligns nicely

with using storage as a “peaker” plant replacement.

Day to day

If the energy and demand differ from day-to-day, then storage system inefficiencies

can arise on any given day when the storage system may not be fully charged or

discharged.

Month to Month

(seasonal)

When generation or demand varies significantly between seasons, to use storage

for smoothing these fluctuations would result in few charge/discharge cycles. For

seasonal storage applications, many analyses assume that the storage is only

discharged once a year, decreasing the capacity factor from the typically assumed

350 charge cycles in a year. This could represent a per MWh energy cost increase

factor of 350 It is generally understood that addressing seasonal intermittency with

storage is prohibitively expensive and is best addressed by peaking gas plants.70

Year to Year Variations

Wind, solar and hydro generation outputs are also affected by annual variations in

weather patterns. This means DER systems need to be designed for an annual

reference and recognize that capacity factors can be affected on a year to year basis.

Table 5 – Types of Renewables Intermittency and Variable Generation

5.1.2 Demand Requirements for DER Solutions

A future - reference demand profile for Ontario has been developed to facilitate the modeling of the

economics of DER solutions and enable an objective comparison of the DER options.

For this analysis, an 8,760-hour annual demand that reflects the 2017 LTEP forecast for 2035 is used. This

demand profile incorporates many assumptions including the degree to which daily demand profiles may

change with the introduction of 2.4 million electric vehicles.

It is assumed that the requirements for the DER systems would reflect a demand profile that could not be

met by the existing and committed resources defined by the 2017 LTEP (as previously discussed in Section

1.0).

70 CPI, 2017


57


Figure 53 illustrates the contributions that existing and committed nuclear, hydro, biomass, and import

supplies provide, and the gap of community baseload and daytime demand that could be filled by DER

solutions. The profile for the committed resource supplies in Figure 53 illustrates that the existing system

capabilities provide some seasonal flexibility and demand management capabilities.

a) Community Demand

Community Demand is defined as the demand that exceeds the capabilities of Ontario’s committed

resources and that is currently being met by variable generation sources such as wind/solar and natural

gas. The capacity of these latter generation resources is covered by contracts that will be gradually

expiring and that will need to be renewed or replaced by 203571. This demand profile could potentially

be met with DER system options.

“Community” is defined as a mix of residential and commercial customers at a common location, e.g. 1000

homes and/or the equivalent in commercial businesses. The associated demand has baseload, daytime,

and peak components. Figure 54 illustrates the community demand profile used for simulation purposes.

The design requirements for the baseload component are fairly straightforward, e.g. constant demand

every day for the entire year. The daytime demand is more challenging as it contains all the intermittency

characteristics previously discussed. The peak component is to be assumed addressed by peaking gas-

fired generation.

71 IESO forecast in 2017 LTEP

Figure 53 – Supply Mix Contributions to Ontario Demand


58


A demand of 2,250 MW is assumed to be met by a future baseload supply. There are several hours in the

year when that baseload supply may be in surplus (as indicated by the weekends when the blue demand

line spikes below the baseload supply reference). This level is used so as to not over-burden the daytime

demand profile with a baseload requirement. The profile of the baseload community demand was derived

from the average output of Ontario’s nuclear fleet between 2011 and 2015. This was deemed reasonable

as the baseload supply is more cost effectively provided by a reliable grid source such as hydro or nuclear

generation.

b) Committed Supply

The committed supply is made up of nuclear72, hydro, biomass, and import/exports from Quebec73. This

supply mix provides some daily supply flexibility as illustrated in Figure 55.

72 Excluding the Pickering Nuclear Station which retires in 2024 73 Only 60% of the impacts of energy trading with Quebec impacts the profile as it is assumed the remainder is comprised of power flows through Ontario to neighboring jurisdictions

Figure 55 – Daily Supply of Committed Resources

Figure 54 – Components of Community Demand


59


The profile above reflects a 3-year average of Ontario supply between 2015 to 2017. for the simulation,

the output for each hour was computed from a running 28-day average to mitigate the effect on the

results that could otherwise be caused by the baseload curtailments that occurred in 2015, 2016 and

2017. This produces a unique 24-hour output profile for each day of the year.

The generation output profile of the committed resources was then subtracted from the total Ontario

demand to create a daily community daytime demand profile unique to each day of the forecast 2035

year. This demand profile is used to assess the capabilities of the different DER options. Figure 56

illustrates the average September profile for daily community demand.

Since the nature of each supply type (wind, solar, nuclear) is different, the initial sizing of the storage

capacity was based on different time frames. The design references set a storage capacity so as to

minimize curtailment and inefficiencies during that reference period. The month of September was used

for solar as demand during this month reflects a 12-hour day in Ontario and it is reasonably sunny. For

wind, March demand was selected as a typical demand month when the wind blows strongest. This

minimizes surplus output and curtailment. Nuclear capacity is based on the month of May, to reflect a low

demand period and to also minimize surplus energy.

Solar-based DER applications can be sized to support communities of 500 homes and up to several

thousand homes. To compare solar-based DER to baseload-supplied DES, a common aggregated demand

pool was selected to represent the demand for 800,000 homes plus the equivalent commercial demand.

This reflects a total annual community daytime demand of 2.5 TWh and a 900 MW peak. To supply this

demand, the three scenarios include: 1,700 MW of solar, 920 MW of wind, and 275 MW of nuclear.

5.2 Solar-Based DER Intermittency Implications

To explore the impacts of intermittency on solar-based DER solutions, the effects of intermittency have

been segregated from those caused by demand fluctuations. To do so, the simulation assumes that all

days of the year have the same daily demand profile. Using the average September daily demand yields

an annual demand of 3.2 TWh. The performance measures impacted by intermittency include:

Figure 56 – Average Daily September Community Demand


60


• Wasted Generation – arises when the solar output exceeds the demand and storage capacity.

• Unused storage – arises when there is insufficient generation to charge the storage or there is

insufficient demand to discharge the storage. This measure impacts on storage capacity factor.

• Need for natural gas-fired backup generation – arises when there is insufficient solar generation

and/or stored energy to meet demand.

The resulting solar-based DER system characteristics are summarized in Table 6. The results of the three-

year simulation of intermittency using actual Ontario solar output are summarized in Table 7. This shows

that 47% of the solar generation will be used directly, 34% will be retrieved from storage, and natural gas

backup will be required to supply 30% of demand.

Table 6 – Solar Constant Demand, System Characteristics

Table 7 – Solar Constant Demand, System Performance

The following subsections first examine the effects of hourly and daily intermittency, and then assess the

seasonal and annual implications.

DER Component Characteristic Value

Total (GWh) 3,248

Peak (MW) 514

Capacity (MW) 1,669

Capacity Factor (%) 19.1%

Capacity (MWh) 4,478

Capacity (Hours) 8.7

Backup Generation Gas Peak (MW) 508

Table 6 - Solar Constant Demand, System Characteristics

Community Daytime

Demand

Solar

Storage

DER Component Performance Metric GWh % of Generation

Used Directly 1,314 47%

Excess 364 13%

Into Storage 1,119 40%

Total Output 2,798 100%

Stored Generation 962 34%

Losses 157 6%

Capacity Factor (%)

GWh % of Used Generation

Backup Generation 971 43%

DER Managed Peak (MW)

Unmanaged Peak (MW)

Managed Peak Reduction %

Capacity Factor (%)

GWh % of Demand

Used Generation 1,314 40%



Wasted Generation 521 16%

Table 7 - Solar Constant Demand, System Performance

61.3%

Backup Generation

508

514

-1%

21.8%

System Totals

Generation

Storage


61


5.2.1 Solar-Based DER Daily Daytime Operation and Daily Intermittency

Solar can have significant variations in output from hour to hour and day-to-day. Figure 57 shows the solar

output for each day in September 201574 and how they vary with respect to the average. The solar-based

DER solution would have the battery capacity sized according to average solar output. The relationship

between the average output and the peak output is an indicator of the capacity factor. The peak output

of a solar array is similar in most months of the year. The average profile for any given month is affected

by how often the peaks are achieved and over how many hours the sun is above the horizon.

As Figure 57 shows, solar output can frequently vary above and below the average and can drop to near

zero for an entire day. It also illustrates that solar requires storage in order to reliably meet demand, even

in relatively sunny months like September. Hour to hour variations of solar output would have minimal

impact on the performance of the DER storage system if the cumulative daily output achieves the average

assumed in designing the system capacity.

When the output is greater than average, surplus solar cannot be stored, as the battery will be fully

charged to its designed capacity before the sun sets. This surplus solar is modelled as “wasted” solar.

When solar output is below average, the battery will not fully charge to its designed capacity and hence

will not be able to discharge sufficient energy after the sun goes down. This leads to unused battery

capacity as well as a need for backup supply to make up for the unavailable solar and/or stored energy.

Figure 58 provides a simplified illustration of how solar intermittency parameters relate to daytime

demand in the month of September. The average solar output is the sum of the used and charging solar

as well as the losses that occur in storage. Since the first draw on the solar output is used to meet demand,

the impacts of intermittency are amplified for the use of storage. Surplus solar refers to solar output that

exceeds demand and which would be used to charge the battery. The DER storage system would be

designed to accommodate the average of this surplus solar. Solar energy losses arise from storage

conversion inefficiencies.

74 Based on September 2015 actuals from IESO data, not curtailed.

Figure 57 – Daily Solar Output Profile


62


The amount of solar output that exceeds the battery capacity is illustrated by the redline in Figure 58.

Excess solar output above the designed storage capacity represents on average 17% of the total solar

output. Since high solar output days occur 50% of the time, 8.5% of total solar output is wasted. Similarly,

the average amount of solar output when there is a shortfall is illustrated by the yellow line in Figure 58.

Low solar output days will occur 50% of the time eliminating 57% of the expected battery charging for

those days, or 28% of total battery capacity. This 8.5% loss increases the cost of solar by 8.5%. At a 72%

battery capacity the cost of storage increases by almost 40% and requires backup generation to supply

that battery storage shortfall.

Figure 59 illustrates how the intermittent solar output for September 2017 integrates with storage and

backup generation over 30 days of a constant average daytime demand profile for September 2035. This

simulation allows for unused stored energy from the previous day to be carried over to the next day.

Figure 59 – Solar and Storage System for September 2017

Figure 58 – Solar and Storage for Daytime Demand Supply


63


Figure 59 demonstrates how the simulation captures the impacts of intermittency. The colors indicate

when storage is being charged and discharged and when backup generation is required to address any

shortfalls in output.

5.2.2 Seasonal and Annual Variations

Figure 60 illustrates how the solar output capacity factor in Ontario varies throughout the year. The

capacity factor has significant seasonal variations, dropping as low as 5% in December from 28% in June.

Lower solar capacity factors in winter result in lower battery capacity usage, because there is not enough

sun to fully charge the battery. and will require more gas backup to cover the shortfall.

The seasonal variation in capacity factors complicates the appropriate sizing of the storage. Over-building

solar capacity will allow better use of the battery in winter and require less gas but will result in unused

solar surplus in the summer months. Over-building the battery capacity will allow more flexibility in the

use of solar in the summer months but will decrease the realized battery capacity factor. Both of these

options lead to increased costs.

Figure 61 illustrates how solar capacity factors can vary from year to year. In 2017 there was significantly

less sun in the spring and fall compared to 2015 and 2016. Yearly variations can be up to 25% of the

seasonal average.

Periods with lower solar output make the batteries less efficient and require more backup generation.

Periods of higher solar output could lead to greater solar waste but an improved battery capacity factor.

There is less choice with the required backup capacity. It must be sized to meet the maximum need

anticipated in order to ensure a reliable supply under foreseeable circumstances.

Figure 60 – Solar Capacity Factor Rolling 28-Day Average


64


5.2.3 Impact of Intermittency on System Parameters Under Constant Demand

The peak backup demand and the associated capacity factor of the backup generation facility are

important components of the total system cost. The manner in which the storage capabilities of a DER

system are operated can be optimized to help minimize the peak gas used and the peak discharge current

of the storage output.

Figure 62 illustrates the needed backup generation for a day in 2035 September when stored energy is

withdrawn and before the backup generation is required. This scenario represents the worst case peak

demand for backup generation.

Figure 63 illustrates a different scenario when the storage is leveraged by averaging the daily gas

requirement as a percentage of the demand in concert with utilized storage and generation over the day.

This approach minimizes the peak demand both for the storage output and for backup generation.

It may not be practical to achieve the full benefits of the “Percent of Demand” approach illustrated in

Figure 63. To do so may require some predictive capability in the controllers and the amount of gas that

can be shifted may be limited by available battery capacity on any given day. For the purpose of the

comparative cost analysis, the average of the required peak backup generation from the two approaches

is used. With this approach, Table 7 shows that the peak backup generation can be reduced from 514 MW

to 508 MW. This would not be a material impact on the peak backup generation demand. However, later

in the analysis, when demand fluctuations are taken into account, this method does materially reduce the

system peak demand for backup generation.

Figure 61 – Solar Output Comparison by Year


65


5.3 Wind-Based DER Implications

This section examines the impacts that wind intermittency has on the performance of a DER solution

assuming a constant demand profile.

To explore the impacts of intermittency on wind-based DER solutions, the effects of intermittency have

been segregated from those caused by demand fluctuations. To do so, all twelve months of the year are

assumed to have the same demand profile as the month of March. Wind generation could potentially

provide either daytime or baseload demand.

Tables 8 and 9 summarize the wind-based DER system characteristics for a daytime and baseload supply

function respectively. The annualized daytime demand is 2.8 TWh and the annualized baseload demand

is 3.2 TWh with both options assuming a 920 MW aggregated wind farm capacity.

Table 8 – Wind Constant Monthly Daytime Demand

Table 9 – Wind Baseload Demand, System Characteristics

The performance results under these two scenarios based on full three-year simulation of actual Ontario

wind output and intermittency are provided in Tables 10 and 11.

DER Component Characteristics Value

Total (GWh) 2,795

Peak (MW) 598

Capacity (MW) 920



Capacity (Hours) 87


Table 8 - Wind Constant Monthly Daytime Demand

Demand

Generation

Storage

DER Component Characteristics Value

Total (GWh) 3,175

Peak (MW) 362

Capacity (MW) 920



Capacity (Hours) 81


Table 9 - Wind Baseload Demand

Demand

Generation

Storage

Figure 62 – Solar and Storage System with Unadjusted Gas Peak

Figure 63 – Solar and Storage System with Gas Averaged as Percent of Demand


66


Table 10 – Wind Constant Monthly Demand, System Performance

Table 11 – Wind Baseload Demand, System Performance

The impacts of intermittency on supplying the daytime demand are that 69% of wind energy will be used

directly, 16% will be utilized from storage, and 20% of the demand will have to be supplied by backup

generation. To provide a baseload generation function, 77% of wind would be used directly, 14% from

storage, and 25% of baseload demand would come from backup generation.

This section discusses the nature of wind intermittency and the criteria required to size the storage facility.

A constant monthly demand is used to illustrate how a wind-based DER solution would respond to

demand. Additionally, the system design parameters required to accommodate seasonal and annual

variations in wind generation output are described and the final simulation results are presented.



Excess 144 6%

Into Storage 659 25%



Losses 230 9%

Capacity Factor (%)




Unmanaged Peak (MW)


Capacity Factor (%)

GWh % of Demand





System Totals

Generation

Storage

3.5%

Backup

Generation

533

556

-4%

11.9%

Table 10 - Wind Constant Monthly Demand, System Performance



Excess 49 2%




Losses 192 7%

Capacity Factor (%)




Unmanaged Peak (MW)


Capacity Factor (%)

GWh % of Demand





System Totals

Backup

Generation

350

362

-3%

26.0%

Table 11 - Wind Baseload Demand, System Performance

Generation

Storage

3.5%


67


5.3.1 Wind Daily intermittency

As with solar, wind output on any given day can vary dramatically. This section illustrates the nature of

intermittency by examining the daily wind output profile, the frequency of wind output magnitude and

the pattern of output over several months, and the degree to which battery size can smooth wind output.

a) Daily Wind Profile

Figure 64 illustrates the output of 920 MW of aggregated wind capacity for several days in March 2015.

Unlike solar, wind patterns do not have a uniquely shaped daily profile. Wind output does not peak or fall

at similar times every day. It can have high or low peaks at any hour of the day.

The average daily output appears to be uniform, with 25% higher output at night than mid-day. However,

the day-to-day output variation can be as high as 800 MW one day, and as low as 20 MW the next.

Variations within the same day can dramatically drop from 700 MW to 100 MW over the course of a few

hours.

b) Frequency of Wind Output Magnitude

Figure 65 illustrates how often wind output magnitude is achieved as a percentage of time in the winter

season.

Assuming a constant baseload demand equal to the average wind output must be supplied, then the

objective of storage would be to shift output from times of high generation to times of low generation.

When wind output is above the average, the energy will be directed to storage. When wind output is

below average, the baseload supply will be achieved by discharging from storage. For the CAES storage

considered for wind, its efficiency losses are 35%.

For CAES system based on wind data for the winter of 2015, the average useable wind output capacity

factor would 35%. Wind output was below this average 48% of the time, and above it 52% of the time.

Figure 64 – Variations in Daily Wind Output


68


For a scenario where a constant baseload supply is provided from the DER system, the chart highlights

two implications:

1. Sizing storage input/output (I/O) circuitry

• Storage I/O would have to be sized in order to accommodate the peak wind output and/ or

peak storage output. Peak storage input is defined by the peak wind output amount that is

above the average. Peak storage output would be the desired average output of the system

when there is no wind, in this case that would be 35% of the installed wind capacity. The “cut-

off” line represents the wind output level above which wind generation injection into storage

is curtailed to reduce the maximum of the storage inflow. This curtailment will only occur ~2%

of the time, representing 0.2% of the energy. Without such curtailment, the storage system

would require 30% more current carrying capacity.

2. Capacity factor for the storage will be at best 73%, increasing costs by 37%

• Capacity factor is defined by the number of hours assumed in the costing or 8 hours a day, or

33% of the time, at the full discharge rate.

• The wind patterns depicted in Figure 64 above show that storage output will on average be

half of the rated output capacity for 48% of the time, which yields a storage capacity factor of

24%. Such a capacity factor represents only 73% of the 33% discharge rate assumed in the

cost estimate. By using the storage for less hours will increase the cost of storage to 1/73% or

137%.

c) Monthly Pattern

Figure 66 illustrates the variability of the wind output during February and March 2015. There is no

discernable pattern, as it varies significantly and erratically on a day-to-day basis and can be very low for

several consecutive days. In Ontario, because of its geographic location, large weather fronts can create

long periods of poor sun and/or wind conditions. These gaps are critical to power system engineers who

must design for infrequent longer duration events in order to meet the loss of load expectation (LOLE) for

periods of critical public demand. From a customer’s perspective, reliable system operation requires

Figure 65 – Winter Wind Output as a % of Capacity


69


infrequent longer duration events that in turn place constraints on the supply mix and/or minimum

storage requirements.

d) Battery Sizing and Wind Smoothing

Several concepts were tested in order to determine a criterion for sizing storage that can best optimize

wind output. Figure 67 illustrates a running average of wind output for 14 days, 21 days and 28 days for

Ontario’s five the highest wind output months, from November to March.

28 days of storage would best smooth the wind intermittent output. The more the wind output is

smoothed minimizes wasted wind energy. However, even with 28-day storage, the very significant

variations that occur from November to January cannot be fully mitigated. To do so would require three

months of storage.

Figure 66 – Hourly Wind Generation

Figure 67 – Storage Size for Smoothing Wind Intermittency


70


5.3.2 Sizing Storage Capacity to Accommodate Wind Output

The primary role of storage is to capture wind energy that exceeds demand at any point in time and then

to discharge that stored energy when the wind output falls below demand. Figure 68 illustrates, for the

March 2015 reference, this interaction between wind output and an ideal storage system (i.e. no wasted

wind output) when supplying the daytime demand profile. While the surplus wind energy is productively

captured, storage discharge occurs intermittently, sometimes several days apart.

Figure 69 illustrates the cumulative balance of stored energy when the above method is applied to a

sequence of the five highest wind energy months. For this analysis, the wind farm was sized to provide

the same amount of energy as would be demanded over that five-month period. It is for this reason that

the storage starts and ends at zero.

Figure 68 – Integrating Wind and Storage

Figure 69 – Cumulative Stored Wind Energy


71


Sizing the battery requires an estimate of how much surplus energy is available for charging and how

much demand will be placed on its output for discharge. The net effect of these two dynamics is how

much the cumulative stored energy could change over a targeted timeframe. This change in cumulative

energy is referred to as the “swing”. Figure 67 suggested that the optimal targeted timeframe is 28 days.

From the cumulative stored energy data, the maximum amount of energy swing in any 28-day period can

be computed. The swing, or change, between the minimum and maximum storage value shows how much

cumulative energy could be put into or withdrawn from storage during that 28-day period. This swing

defines the capacity of storage that would be required to smooth a 28-day wind output. Figure 70

illustrates the running 28-day swing of storage for two scenarios: (1) supplying the daytime demand; and

(2) supplying a constant baseload output.

For the daytime function, the average swing is approximately 50 MWh for a 1 MW wind farm. A smaller

storage capacity would be required for a baseload function versus a daytime supply function. The

baseload option would require approximately 42 MWh of storage for a 1 MW wind farm. However, by

sizing the storage for the average swing, the storage will not be able fully accommodate the surplus

energy. Surplus energy that cannot be stored would be wasted. Conversely, for periods when the swing

is less than then average, the storage would be underutilized. Sizing the storage capacity based on the

average achieves an approximately 50-50 balance between times of wasted energy and times of

underutilized storage capacity.

Alternatively, storage could be sized to the minimum swing, thereby maximizing the utilization of the

storage capacity but also maximizing wasted wind. Since unused storage capacity is more expensive than

wind energy, the system has been sized for simulation purposes halfway between the average storage

and the minimum storage.

Figure 70 – Maximum 28-Day Stored Energy Swing


72


5.3.3 Results of Storage Sizing for Design Period

Table 12 summarizes the key parameters of a wind-based DER system designed to optimize the use of

wind energy for the five highest wind output winter months. Actual wind output data was used for all five

months. To bring out the effects of intermittency, the March reference demand was replicated each

month. Table 13 summarizes DER performance measures in terms of generation, storage, and backup

generation.

Table 12 – Wind Constant Monthly Demand for Winter, System Characteristics

Table 13 – Wind Constant Monthly Demand for Winter, System Performance

Sixty three percent (63%) of direct wind generation output and 18% of the wind energy from storage is

used to meet demand for a total utilization of the wind output of 81%. The remainder is either lost or

wasted.

Excess generation refers to the wind output that is above demand and cannot be stored because the

storage is at its maximum capacity at the time. The amount of this wasted excess generation for this

period is 9%.

This simulation assumes that wind is stored using CAES, which has a round trip efficiency of 65%, or a loss

factor of 35%. Lost wind energy due to storage inefficiencies is thus 9% of total wind output.

Backup generation in this scenario is assumed to be natural gas-fired generation. The total gas output

required is very small representing 3% of demand.

DER Component Performance Metrics Average

Total (GWh) 1,159

Peak (MW) 598

Capacity (MW) 920



Capacity (Hours) 87


Demand

Generation

Storage

Table 12 - Constant Monthly Demand for Winter


Used Directly 884 63%

Excess 131 9%




Losses 132 9%

Capacity Factor (%)


Backup Generation 29 2.6%


Capacity Factor (%)

GWh % of Demand

Used Generation 884 76%




Table 13 - Wind Constant Monthly Demand for Winter, System Performance

Generation

Storage

5.3%

374

2.1%

Backup

Generation

System Totals


73


For the winter 2015 scenario a gas plant with a peak capacity of 374 MW is required to meet the shortfall

requirements of the 920 MW wind farm. This gas plant will also need to supply a total of 29 GWh of output

over the entire season.

These results suggest this approach is an effective way to optimizing the use of the wind energy. However,

the storage capacity factor is approximately 5%, which will multiply the unit cost of storage by a factor of

almost 20, pushing the cost of energy delivered from storage to over $1,000/MWh.

5.3.4 Seasonal and Annual Fluctuations

Wind energy does not sustain a high output all year long, and these output levels vary from year to year.

Figure 71 shows the wind output in Ontario for the three years 2015 to 2017. Generally, wind generation

follows a seasonal profile, with the peak coming in winter and a trough in the summer. However, the

output could vary by as much as 40% from the seasonal average in any given period.

Periods of lower wind output makes storage less efficient and requires more backup generation. Periods

of higher wind output should result in improved battery capacity factors but will have higher amounts of

excess generation.

The backup capacity needs to be sized for maximum need as the plant size cannot be changed, and this

reduces the capacity factor of the backup system and further increases costs.

5.3.5 Impact of Intermittency on System Parameters under Constant Demand

Tables 14 and 15 summarize the results of a three-year simulation of intermittency using actual Ontario

wind output with an indication of the sensitivity of the results to year-over-year intermittency variations.

Figure 71 – Wind Output Comparison by Year


74


Table 14 – Wind Constant Monthly Demand, 3 Year Comparison

Table 15 – Wind Baseload Demand, 3 Year Comparison

To meet the daytime demand, 69% of wind energy will be used directly, 16% will come from storage, and

20% of demand will be supplied by backup generation. For the baseload function, 77% of wind would be

used directly; 14% from storage; and 25% of demand will be supplied from backup generation.

DER system performance varies year-over-year. For the daytime demand scenario, the annual differences

can change wasted excess energy by +33%/-55% variation, the capacity factor for storage by +8%/-7%,

Average

DER Component Performance Metrics 2015-2017 High Low

% of Gen.

Generator CF 32.3% 3.9% -2.1%

Used Generation 69% 0.6% -0.4%

Excess Generation 5% 33.6% -55.0%

Into Storage 25% 10.4% -6.2%

Stored Generation 16% 10.3% -6.2%

Storage Losses 9% 10.5% -6.2%

Storage CF 3.5% 4.2% -8.1%

% of Used Gen.

Needed Gas 25% 15.0% -16.2%

Total Gas Needed (GWh) 555 12.2% -13.9%

Peak Gas Capacity (MW) 533 - -

Backup Generation CF 11.9% 12.2% -13.9%

% of Dem.



Needed Gas 20% 12.1% -13.9%

Wasted Generation 13% 9.0% -16.5%

Table 14 - Wind Constant Monthly Demand, 3 Year Comparison

Generator

Storage

Backup

Generation

System Totals

% Sensitivity

Average

DER Component Performance Metrics 2015-2017 High Low

% of Gen.

Generator CF 32.3% 3.9% -2.1%


Excess Generation 2% 68.1% -73.7%

Into Storage 21% 3.2% -3.8%


Storage Losses 7% 3.3% -3.9%

Storage CF 3.5% 11.4% -10.1%

% of Used Gen.

Needed Gas 34% 7.1% -11.2%

Total Gas Needed (GWh) 797 6.1% -9.4%

Peak Gas Capacity (MW) 350 - -

Backup Generation CF 26.0% 6.1% -9.4%

% of Dem.



Needed Gas 25% 5.9% -9.3%

Wasted Generation 8% 17.7% -14.1%

Table 15 - Wind Baseload Demand, 3 Year Comparison

System Totals

Generator

Storage

Backup

Generation

% Sensitivity


75


and the amount of required backup generation by +15%/-16%. The required peak backup generation is

consistent across the three years.

5.4 Demand Fluctuation Implications

Demand fluctuations present challenges for all DER options as they impact the efficiency of these options

much like the intermittency of renewables. The impacts of intermittency and demand fluctuations on the

DER option are illustrated in Figure 72.

The characteristics of a system based on the average output that assumes no intermittency, as described

in Section 4.4, is contrasted against the impacts of the intermittent nature of renewables and the reality

that demand also fluctuates. The primary system measures brought out in the comparison is the wasted

generation and the need for backup generation. Wasted generation refers to both losses in the storage

system as well as any surplus generation that would get curtailed. As such, even the perfect no-

intermittency case has wasted energy in the form of storage losses.

The “No Intermittency” scenario refers to an ideal case where average output and demand are used as

inputs and are assumed to be constant for an entire year. The “Renewables Intermittency” scenario

considers the impact on DER performance characteristics of daily variations in solar and wind output with

demand patterns held constant throughout the year. Intermittency wastes energy, reduces storage

utilization, and requires backup generation. Unlike intermittent renewables, nuclear baseload-supplied

DES is not a variable generation resource and hence has no impact due to intermittency.

Fluctuations in demand further increase waste and backup generation but can increase the use of storage.

This section looks at the sources of demand fluctuation and how this results in performance degradations

of the DER system options.

Figure 72 – Impact of Intermittency and Demand on DER


76


5.4.1 Understanding Demand Fluctuation

Demand has its own fluctuations: month-to-month (seasonal); daily; and weekly. Each are described

below.

Figure 73 illustrates the annual profile for Community Daytime demand used for this analysis. While the

prevailing view is that demand peaks in the summer, the average daily demand is noticeably higher in the

winter, averaging approximately 4,000 MW per hour.

For the rest of the year, the average daily demand increases from ~2,200 MW in the spring to ~3,000 MW

in the summer and then falls back to ~2,500 MW in the late fall. This means that any generation built for

the winter would likely be underutilized during the rest of the year. However, this higher winter demand

fits well with the design assumptions for the wind-based DER option. Winter was used as the reference

demand case because it is the time when both demand and wind are at their highest. Other periods of

the year were used to size both the solar-based DER (September) and nuclear baseload-supplied DES

options (May).

Higher than average daily winter demand is influenced by the swing between the low and high demand

that occurs in a day. The average daily profiles for winter and summer are illustrated in Figures 74 and 75.

Figure 73 – Annual Community Daytime Demand Profile

Figure 74 – Average Daily Winter Community Daytime Demand Profile

Figure 75 – Average Daily Summer Community Daytime Demand Profile


77


The difference or demand swing from the night time lows to the daytime highs are more pronounced in

summer than in winter. The summer nighttime low in demand is over 70% lower than the daytime peak,

compared to only a 45% difference in winter. This means that storage has to work harder in the summer.

This circumstance should favor solar-based DER systems as the solar generation also comes on during the

day with storage providing less of the daytime demand as compared to the wind-based DER or nuclear

baseload-supplied options.

Demand on a day-to-day basis varies significantly at all times in the year. Figures 76, 77, and 78 illustrate

the range of hourly demand that can occur over the day in the design reference months of March (wind),

May (nuclear), and September (solar). This range translates into demand uncertainty day to day.

March has the least amount of demand variability with demand uncertainty of 4,000 MW at an almost

constant level of throughout the day. May nighttime demand variations have the smallest uncertainty of

approximately 2,500 MW, but with daytime demand swings of 5,500 MW. September has the widest

possible daily demand variations of 4,000 MW nighttime and 7,800 MW daytime. These variations show

that even in the designed months, demand fluctuations will impact wasted energy, unused storage

capacity, and the need for backup generation similar to intermittent renewables.

Weekend demand is another factor that contributes to the wide variations in average daytime demand.

Over the year, weekend demand is on average 40% less than weekday demand as illustrated in Figure 79.

All DER options will in general have lower usage factors on weekends. It is also evident that energy use is

higher earlier in the work week.

Figure 76 – Daily March Demand Profile Figure 77 – Daily May Demand Profile

Figure 78 – Daily September Demand Profile


78


5.4.2 Annual Demand Variations

The annual variations in demand are another important factor for the designing of DER solutions that will

be expected to operated for many decades. Figure 80 shows how the annual demand profile varied

significantly during the period from 2015 to 2017.

The year-to-year differences in demand are due to uncontrollable factors such as the weather. Potential

variations can be highlighted by comparing the demand for the months of February and September

between 2015 and 2017. Demand in February 2015 was 30% higher than in 2017, while September 2015

demand was 35% higher than in 2017. Such significant variations will impact DER system performance i.e.,

wasted energy, storage capacity and backup generation requirement. The impact of annual demand

variations has not been incorporated into the analyses of this report.

Figure 79 – Community Daytime Weekly Demand Profile

Figure 80 – Yearly Ontario Demand Profile Comparison


79


5.4.3 Implications of Demand Fluctuations on DER options

The fluctuations in demand present different challenges for each DER option.

5.4.3.1 Implications of Demand on Solar-Based DER

As discussed earlier, the solar-based DER model was sized to optimally supply the demand from an

average September day. Figure 81 shows how fluctuations in demand would interact with solar

intermittency and impact the performance of a DER system during the month of September.

Overlaying demand and supply intermittency demonstrates that there is insufficient solar energy output

during periods of high demand or low solar activity. Conversely, during periods of low demand, solar

energy gets wasted and the storage never fully discharges.

Table 16 summarizes the results of full year simulations of the solar-based DER options and how

performance varied between constant and real demand fluctuation scenarios.

Figure 81 – Solar and Storage System for September 2017


80


Table 16 – Comparison of Solar Constant Demand and Full Year Demand Cases

The impact of demand fluctuations on a solar-based DER system is best understood by visualizing the

interplay between demand and solar output over a full year. This visualization is provided in Figure 82

which highlights several factors:

1. Backup generation is primarily needed in the winter months but also late fall, coinciding with low

solar capacity factors.

2. Solar output can be strong in any season, even January. However, more high output days

increasingly occur in the spring and summer.

3. Summer peak demand does not correspond with peak solar production.

4. The majority of wasted solar occurs in the low demand spring season.

The performance of the DER system is impacted by the size of the solar panels and the capacity of the

storage. While some design approaches size the system to produce the energy required to meet demand,

this analysis sized the system based on the month of September, which has 12-hour days of average

sunshine and demand.

While future work could be conducted to optimize the design parameters beyond the reference case

assumptions used here. Figure 82 suggests that optimization may only have a marginal impact on the

net results. Increasing the size of the solar panel may provide some additional energy during periods of

low solar output to increase the storage capacity factor and to offset backup gas generation. However,

this would increase wasted energy to a far greater degree in spring and summer due to the ratio of the

capacity factors between the seasons. Increasing the storage size is unlikely to materially reduce the

need for backup generation as there is no solar energy available when backup generation is required

(e.g. in winter). The marginal benefit would likely be offset by the cost of the larger and costlier battery

capacities.

DER Component Performance Metric Constand Demand Full Year Demand % Change

Used Directly 47% 35% -25%

Excess 13% 31% 138%

Into Storage 40% 34% -15%

Total Output 100% 100% 0%

Stored Generation 34% 29% -15%

Losses 6% 5% -15%

Capacity Factor 61.3% 52.1% -15%

Backup Generation 43% 43% 0%

DER Managed Peak (MW) 508 722 42%

Unmanaged Peak (MW) 514 865 68%

Managed Peak Reduction % -1% -17% -

Capacity Factor (%) 21.8% 12.2% -44%

Used Generation 40% 38% -6%

Stored Generation 30% 32% 8%

Backup Generation 30% 30% 1%

Wasted Generation 16% 39% 144%

System Totals

Table 16 - Comparison of Solar Constant Demand and Full Year Demand Cases

Generation

Storage

Backup

Generation


81


Figu

re 8

2 –

Sola

r-b

ase

d D

ER S

yste

m B

eh

avio

r


82


5.4.3.2 Implications of Demand on Wind-Based DER

As previously discussed, the wind-based DER model was sized to optimally supply the average March

demand.

Demand fluctuations significantly impact storage sizing. Figure 83 illustrates the sizing criteria for storage

under real demand conditions and contrasts that with the same criteria discussed earlier based on the

constant demand scenario. The average swing in the cumulative energy in storage is about 80 MWh.

Averaging this with the minimum swing gives a storage size required to accommodate actual demand –

53 MWh for 1 MW wind farm. This is larger than the 42 MWh previously identified as required by the

constant demand scenario. The cumulative stored energy swing shown in Figure 83 arises from excessive

surplus generation in February. A larger storage capacity is required to shift that energy to March.

An assessment was conducted to determine the sensitivity of DER performance to storage size. Table 17

shows how 28-day, 21-day, and 10-14 days of storage impact the performance parameters for the 920

MW wind DER system.

Figure 83 – Required Demand Storage Outflow


83


Table 17 – Wind Storage Size Sensitivity Analysis

The results in Table 17 suggest that a smaller storage size could meet the assumed demand. Halving the

storage capacity would only increase excess generation by 25% (from 8% to 10% of energy wasted) and

need for gas-fired generation by 10% (from 20% to 22% of used generation). While the amount of energy

stored declines to 9%, the storage capacity factor increases by 85% (from 2.8% to 5.2%). The lower storage

capacity and improved capacity factors could reduce the expected cost of storage by about a factor of 4.

For the full year simulation, a 10-14 days storage capacity was assumed. Referring back to Figure 67, this

is the period required to satisfy the demand in late March. These results demonstrate the impracticality

of designing Ontario’s future electricity system based upon 100% solar and/or wind with storage as the

LOLE requirements will result in uneconomic amounts of storage.

Figure 84 visually depicts the results of the full year wind simulation and illustrates several factors:

1. Wind output is well matched and balanced to demand in the winter months

2. Most of the excess generation in the model occurs in the spring and late fall when demand is low.

3. The need for gas generation is primarily during the high demand periods of late summer when

wind output is lowest.

The pathways to further system performance optimization are not apparent from the results observed.

Demand Scenario Constant Month

Storage (Days) 28 21 10-14 28

Storage Size (GWh) 48 35 24 35

Used Generation 64% 64% 64% 69%

Excess Generation 8% 9% 10% 5%

Into Storage 29% 27% 26% 25%

Stored Generation 19% 18% 17% 16%

Storage Losses 10% 10% 9% 9%

Storage CF 2.8% 3.7% 5.2% 3.5%

Needed Gas (GWh) 426 445 463 555

Needed Gas 20% 21% 22% 25%

DER Managed Peak (MW) 533

Unmanaged Peak (MW) 556

Managed Peak Reduction % -4.0%

Backup Generation CF 5.6% 5.9% 6.1% 11.9%

Used Generation 65% 65% 65% 65%

Stored Generation 19% 18% 17% 15%

Needed Gas 17% 17% 18% 20%

Wasted Generation 18% 19% 20% 13%

% of Generation

% of Used Generation

% of Demand

System Totals

Storage Size

Inter

Case

Recommended

Case Reference Case

Full Year

Table 17 - Wind Storage Size Sensitivity Analysis

862

866

-0.5%

Backup Generation

Scenario Name

Demand

Case

Generation

Storage

DER Component


84


Figu

re 8

4 –

Win

d-b

ase

d D

ER S

yste

m B

eh

avio

r


85


5.4.3.3 Implications of Demand on Nuclear Baseload-Supplied DES

The nuclear baseload-supplied DES model was sized to optimally supply the average May demand. Figure

85 shows how the fluctuations in demand interact with the nuclear baseload supply for the month of May.

The overlay of demand and supply intermittency shows that when periods of high daytime demand occur,

there was insufficient surplus nuclear energy at night to sufficiently charge the batteries resulting in the

need for backup generation from other sources. Conversely, during periods of low demand, which for May

was invariably weekends, nuclear energy is curtailed as the small battery size became fully charged.

Table 18 summarizes the performance of the nuclear baseload-supplied DES option for the constant

demand and the full year real demand fluctuation scenarios. Under the constant demand scenario,

nuclear baseload characteristics are a good fit with the optimal operation of a DES system. The benefits

include: no wasted nuclear output; a 100% battery storage capacity factor; and, no gas generation backup

requirement. Clearly, this is not the case for a real demand scenario where ten percent of nuclear output

is wasted, battery capacity factors decline to 28.6% and backup gas generation is needed to supply 24%

of demand.

Figure 85 – Nuclear and Storage System for May


86


Table 18 – Comparison of Nuclear Constant Demand and Full Year Demand Cases

Figure 86 visualizes the results of the full year nuclear simulation and illustrates several factors:

1. Most backup generation is required in the winter months, coinciding with high average demand75,

but is also present during the peak summer days.

2. Backup generation is also required in early fall when nuclear energy is at its lowest.

3. Most nuclear output is wasted in the low demand spring season but with dips occurring around

the July long weekend and at Christmas. Increased storage capacity could help manage these

circumstances, however, only for 6 months of the year.

A sensitivity assessment was conducted to identify how the system performance parameters would vary

with storage size. The results are summarized in Table 19.

Table 19 – Sensitivity Analysis Results for Nuclear-Based DER Storage Size

75 Note that the 91.5% nuclear capacity factor was optimized for peaks in winter and summer to reflect lower output levels in the spring and fall. Between 2011 and 2015, the performance of Ontario’s nuclear fleet on average was able to achieve such patterns.

DER Component Performance Metric Constand Demand Full Year Demand % Change

Used Directly 95% 84% -12%

Excess 0% 10% -

Into Storage 5% 6% 15%

Total Output 100% 100% 0%

Stored Generation 5% 5% 15%

Losses 1% 1% 15%

Capacity Factor 100% 28.6% -71%

Backup Generation 0% 32% -

DER Managed Peak (MW) 0 573 -

Unmanaged Peak (MW) 0 621 -

Managed Peak Reduction % - -8% -

Capacity Factor (%) - 12.3% -

Used Generation 95% 71% -25%

Stored Generation 5% 4% -3%

Backup Generation 0% 24% -

Wasted Generation 1% 9% 1163%

Generation

Storage

Backup

Generation

System Totals

Table 18 - Comparison of Nuclear Constant Demand and Full Year Demand Cases

% Difference from Reference -67% -33% 0%* +33% +300% +1233%

Storage Size (MWh) 378 756 1,134 1,512 4,537 15,125

Storage Size (Hours) 1.0 1.9 2.9 3.9 11.7 38.9

Wasted Nuclear (%) 13.5% 11.9% 11.0% 10.2% 6.4% 2.3%

Battery Capacity (%) 51.0% 36.2% 28.6% 23.7% 11.9% 4.9%

Backup Generation (GWh) 665 637 620 607 543 473

Backup Generation Peak (MW) 573 573 573 573 573 573

*Reference Design Case

Table 19 - Sensitivity Analysis Results for Nuclear-Based DER Storage Size


87


The nominal design storage size is 1,134 MWh or 2.9 hours of storage. Varying the amount of storage

capacity by +/- 33% results in the following effects:

• The capacity factor of the storage changes to 36.2% with smaller storage (a 27% improvement) or

to 23.7% (a degradation of 17%) with larger storage.

• Underutilized nuclear production increases to 11.9% with smaller storage (a +9% degradation) or

to 10.2% (a 7% improvement) with larger storage. This is not a material change.

• The need for backup generations changes by +2.7%/- 2.1%, also not a material change. The need

for peak gas-fired backup is unaffected.

Most likely the potential cost changes for the backup generation requirements and underutilized nuclear

production are more modest than those for altering the storage size. Reducing the storage size by 66% to

achieve the 50% capacity factor is not likely to be as optimal as the impacts on underutilized nuclear and

the requirement for backup generation become larger. For the purposes of this assessment, 2.9 hours of

storage was assumed to be adequate.

Further work may be warranted to choose an optimal balance, including the benefits of increasing the

nuclear capacity, and must consider the costs as a criterion. Costs are addressed in the next section.


88


Figu

re 8

6 –

Nu

cle

ar B

ase

load

-Su

pp

lied

DES

Be

hav

ior


89


5.5 Summary Implications: Comparative Performance of DER Options

The results of the supply and demand analysis for the DER options are summarized and compared in this

section from two perspectives:

1. DER System Design Characteristics

2. DER Option Performance Indicators

The performance measures observed for the nuclear solution suggest it will be the most cost effective.

5.5.1 DER System Design Characteristics

Table 20 summarizes the system design characteristics for the solar-based, wind-based and nuclear

baseload-supplied DER/DES options that have been developed to supply the projected 2035 LTEP demand

for an aggregation of communities.

Table 20 – Full Year Demand System Characteristics for DER Options

As the nuclear and wind generation options are grid-based and necessarily larger scale, an aggregated

community daytime demand was modelled to reflect the combination of 800,000 homes and a

commercial equivalent demand. This demand is over and above that which is supplied by Ontario’s

committed resources for 2035. Total annual demand would approximate 2,500 GWhs, with an annual

peak demand of ~890 MW.

This demand could be met with 1670 MW of solar capacity coupled to 4500 MWhs or 11.5 hours of

storage. Since wind has a higher capacity factor than solar, it requires a lower installed capacity of 920

MW. Nuclear with its 91.5% capacity factor can meet the demand with 270 MW of capacity. The wind-

based DER system includes 62 hours of storage and nuclear only requires 3 hours of storage.

The wind solution has the highest backup generation capacity requirement at 860 MW, almost 60% more

than the 575 MW required by the nuclear baseload-supplied DES option.

5.5.2 DER Option Performance Indicators

Three years of supply intermittency have been modeled for the wind and solar options. The same 2015

demand projection was used for all three years.

DER Component Characteristic Wind Solar Nuclear

Total (GWh) 2,562 2,562 2,562

Peak (MW) 893 893 893

Capacity (MW) 920 1,669 272

Capacity Factor (%) 32.3% 19.1% 91.5%

Capacity (MWh) 24,350 4,478 1,134

Capacity (Hours) 62 11.5 2.9

Backup Generation Gas Peak (MW) 862 722 573

Table 20 - Full Year Demand System Characteristics for DER Options

Community Daytime

Demand

Generation

Storage


90


Figure 87 illustrates for each DER option how each energy source contributes to meeting the community

demand. The solar option requires the most gas backup generation and produces the greatest amount of

wasted energy. Nuclear requires the least amount of storage and wastes the least energy. Wind is in

between the nuclear and solar options with respect to the use of storage and the amount of wasted

energy yet has the lowest need for backup generation. As Table 20 indicates, the wind-based DER has the

highest need for storage and backup capacity suggesting the costs of these will be greater for wind.

The performance characteristics of the DER/DES options are summarized in Table 21 and suggest that the

nuclear baseload-supplied DES option may have the most efficient use of resources.

Figure 87 – DER Generation Required for Community Daytime Demand


91


Table 21 – Full Year Demand Performance Metrics for DER Options

The DER options vary on many of the following key parameters:

• Utilized and stored generation

o After storage losses, nuclear baseload-supplied DES has the highest energy utilization rate

at 89%, either directly or discharged from storage. This is an important a balancing factor

for the DES system LCOE as nuclear generation has high capital costs.

o Wind has the next highest use of generation at 80%. Solar is the least efficient option with

only 64% of solar output used either directly or via storage.

o When considering excess energy and losses in storage, only 11% of nuclear generation is

unused, starkly different than the 19% of wind and 35% of solar.

o Wind, the lowest cost generation option, has the greatest storage losses. The wind option

has the least efficient storage due to the use of CAES. Even though CAES is one third the

cost of the Li-ion battery storage, it has a 35% loss factor on the energy storage cycle.

Curtailing wind generation to save storage costs could provide a balancing factor to the

DER system LCOE and economics.

• Storage Capacity Factor

o Solar, with a storage capacity factor of over 50%, makes the best use of its storage

capacity. Since Li-ion batteries are the most expensive component of the solar-based DER

solution, high capacity factors are important.

o Nuclear has a lower storage capacity factor than solar, but also requires much less storage

capacity.

o Despite 26% of wind energy being cycled through storage, wind-based DER storage

capacity factors are extremely low due to the large numbers of hours of storage required

to capture the wind energy.

• Backup Generation Requirement

GWh % of Generation GWh % of Generation GWh % of Generation

Used Directly 1,661 63% 979 35% 1,829 84%

Excess 271 10% 868 31% 221 10%

Into Storage 684 26% 950 34% 132 6%

Total Output 2,617 100% 2,798 100% 2,182 100%

Stored Generation 446 17% 817 29% 113 5%

Losses 239 9% 133 5% 18 0.8%

Capacity Factor (%)

GWh % of Used Generation GWh % of Used Generation GWh % of Used Generation

Backup Generation 463 28% 770 43% 620 32%


Unmanaged Peak (MW)


Capacity Factor (%)

GWh % of Demand GWh % of Demand GWh % of Demand

Used Generation 1,661 65% 979 38% 1,829 71%

Stored Generation 446 17% 817 32% 113 4%

Backup Generation 463 18% 770 30% 620 24%

Wasted Generation 510 20% 1001 39% 239 9%

Wind Solar NuclearDER Component

Backup

Generation

Generation

Storage

System Totals

-17%

12.2%

Performance Metric

5.2%

862

866

-0.45%

6.1%

28.6%

573

621

-8%

12.3%

52.1%

722

865

Table 21 - Full Year Demand Performance Metrics for DER Options


92


o Wind requires the least amount of backup generation but has the highest peak generation

requirement to cover periods with no wind or stored energy. As a result, the capacity

factor of the backup generation is only 6%. This is half the capacity factor of the solar-

based DER or nuclear baseload-supplied DES options, which are similar at 12%.

o The solar option requires 25% more total backup generation than the nuclear option.

o Wind requires backup generation to provide 18% of the community demand and the

nuclear and solar options 24% and 30%, respectively.


93


6.0 Ontario’s Geographical Challenge

This section looks at the cost of implementing the DER/DES options in Ontario from a total LCOE

perspective. The five major findings are:

1. Intermittency of renewables and demand fluctuations can each increase the cost of DER/DES solutions

as shown in Figure 88:

a. Renewables intermittency increases the expected costs of DER systems when compared to the

ideal model where average renewable outputs are assumed against a constant demand. The

impact of wind intermittency is substantial at a factor of 5 cost increase. Solar has a 5% cost

increase. No intermittency is assumed for nuclear.

b. When demand fluctuations are included in the simulations, solar-based DER costs increase a

further 30%. Surprisingly, the cost of wind-based DER drops when demand fluctuations are

reflected. The nuclear baseload-supplied DES costs increase by almost 60%.

2. Conventional nuclear baseload-supplied DES will be almost 40% less expensive than solar-based DER

options and less than half the cost of wind based DER options.

3. Figure 89 shows that Ontario DER systems will cost approximately 10-15% more than the same

systems installed in most locations in the U.S.

4. Renewable-based DER options could be twice the cost of SMR-based76 or carbon capture options.

a. The forecast cost of a CCGT solution fitted with CCS technology is similar to an SMR-based DES

solution.

5. Wind solutions are considerably more expensive, over twice the cost of the nuclear baseload-supplied

DES option.

76 Assumes SMR generation LCOE of approximately CAD $75/MWh based on EIRP. LCOE DER comparisons based on substituting SMR LCOE for conventional nuclear with all other assumptions remaining the same.

Figure 88 – Intermittency Implications on LCOE of Ontario DER Options


94


The cost implications are based on several factors. The following sub-sections examine:

• The costs to purchase and install DER systems in Ontario;

• Total system implementation costs;

• The impacts of renewables intermittency and demand fluctuations on Ontario LCOEs;

• U.S. geography driven competitive disadvantages for Ontario DER systems.

6.1 DER Component LCOEs for Ontario

Ontario’s DER cost estimates are derived from the costs of the individual system components as discussed

previously based on available U.S. average costs. Different capital cost and capacity characteristics exist

in the U.S. Several cost adjustments have been made to highlight the differences between the two

jurisdictions. These include:

1. Exchange rate, applied to 60% of the cost of solar and wind

2. Cost premium for building in Ontario based on EIA estimates

3. Average Ontario capacity factor differences for community-based solutions

The net impact of these cost adjustment factors on solar-based DER and wind-based DER LCOEs are

illustrated in Figures 90 and 91.

The estimates are compiled from the 2030 cost forecasts described earlier, a community scale solar LCOE

of US $77/MWh including Tx/Dx connection cost becomes a LCOE of CAD $120/MWh for an Ontario

community scale solar installation. Similarly, the 2030 forecast cost of US $42/MWh for grid-scale wind

installations rises to CAD $78/MWh in Ontario.

Figure 89 – Ontario vs. U.S. LCOE Comparison Supplying Daytime Demand


95


6.1.1 Exchange Rate Assumptions

Exchange rate assumptions have been made to facilitate cost comparisons with Ontario installations in

Canadian dollars. There are two components to the exchange rate impacts:

1. The exchange rate itself

2. Imported content of the energy solutions being installed

An exchange rate of 15%, or more specifically 1.15 CAD for 1 USD, is assumed for this analysis. This

exchange rate is based on the average exchange rate between the U.S. and Canadian dollars for the 30

years ending in 201577 and is conservatively low.

• The FAO predicts an average long-term exchange rate of 16.3% out to 2050 but predicts it will

remain above 21% between 2020 and 2030, the period of interest to this study.78

• A low exchange rate assumption depresses what the costs could be in Canadian dollars. This

means the cost represented here is likely to be lower than can be expected, particularly for

renewables and storage.

The exchange rate is applied only to the imported content of the energy options. The rationale is that

while exchange rates have fluctuated significantly since the 2008 recession, inflation in the two countries

has been similar suggesting that cost factors have not changed since the currencies were at par several

years ago. 79 As a result, the ability to construct DER solutions in Ontario is assumed to be similar to the

economically equivalent regions of the U.S., except when the component content is purchased from

abroad. When component content is imported, the exchange rate is applied.

Table 22 summarizes the imported content assumptions that have been applied to the various

technologies. The assumption for renewables reflects the earlier unfavourable World Trade Organization

(WTO) ruling on Ontario’s 50% domestic content Feed-in-Tariff (FIT) requirement. This is a conservative

77 Strapolec, Pickering study referring to OANDA. CAD/USD Historical Exchange Rates. Rates retrieved from Oanda 2018 78 FAO, 2017 79 Inflation.eu, 2018

Figure 90 – Solar (Community-Scale) Ontario LCOE Components

Figure 91 – Wind (Grid-Scale) Ontario LCOE Components


96


assumption, which could understate the solar and wind costs. Batteries are assumed to have imported

content similar to that of wind and solar. The imported content costs for nuclear, CAES, pumped hydro,

and Tx are assumed to be low given the extensive supply chain in Canada for these established

technologies. Natural-gas fired generation has a high imported value due to the cost of the imported fuel,

in particular from U.S. sources.

Table 22 – Imported Content (%)

6.1.2 Installation Cost Premium for Ontario

Many reports of low cost of renewables are from regions that can be characterized as low cost economic

zones. For example, the forecast costs in the Leidos report reflects costs in the gulf coast region, where

capital costs are cheaper and output is more reliable. Leidos explicitly states that adjustment factors

should be applied for other jurisdictions.

The EIA has investigated the differing costs of renewable installations across the U.S.80. These differences,

reflecting the relative cost multipliers for solar and wind installations in the U.S. defined electrical regions

are summarized in Table 23. The EIA also provided the generic costs factors multipliers that the Leidos

study used to produced future estimates for the EIA81. The Leidos parameters are assumed to be generic

to all electrical capital projects.

80 EIA, 2017 81 Leidos, 2016

Solar 60%

Wind 60%

Nuclear 25%

SMR 50%

Tx & Contingency 25%

Li-ion Storage 60%

Pumped Storage 25%

CAES 25%

Backup Gas CCGT w/ CCS* 90%

* 90% is due to considering imported fuel

Storage

Generation

Table 22 - Imported Content (%)


97


Table 23 – Regional Cost Multiplication Factors

The location of the regions in Table 23 are illustrated in Figure 92.82

82 Map sourced from EIA Assumptions to the AEO, 2017

Region Number Region Name Leidos EIA Solar EIA Wind

1 ERCT 0.94 1.08 0.78

2 FRCC 1 0.87 N/A

3 MROE 1.21 1.03 1.06

4 MROW 1 0.93 0.88

5 NEWE 1.05 1.20 1.19

6 NYCW 1.38 1.60 N/A

7 NYLI 1.32 1.02 1.08

8 NYUP 1.05 0.97 1.08

9 RFCE 1.06 1.13 1.08

10 MICHIGAN 1 1.48 1.06

11 RFCW 0.95 0.98 1.06

12 SRDA 0.97 0.93 1.15

13 SRGW 1.05 0.81 1.06

14 SOUTHERN 0.98 0.82 1.15

15 TVA 0.95 0.69 1.15

16 SRVC 0.92 0.86 1.15

17 SPNO 1 0.72 0.73

18 SPSO 0.8 0.93 0.73

19 DSW 1.03 1.10 0.95

20 NP15 1.12 1.16 0.96

21 NWPP 1.05 0.73 0.95

22 ROCKIES 1 0.95 0.73

Average 1.04 1.00 1.00

Average of ON similar regions 1.07 1.16 1.10

Gulf coast avg (SRDA & ERCT) 0.96 1.01 0.97

Ontario to U.S. Avg ratio 103% 116% 110%

Table 23 - Regional Cost Multiplication Factors


98


It is assumed that Ontario’s cost reference would be similar to that of the northeastern U.S. regions, e.g.

MORE (Wisconsin), NEWE (New England), NYUP (Upstate New York), RFCE (Pennsylvania), and RFCM

(Michigan). The cost of solar and wind installations in Ontario’s neighbouring jurisdictions is relatively

expensive. The rest of the eastern U.S. is less so. RFCW is not included in the Ontario average as they are

known to have significantly lower economic costs than Ontario. Based on the EIA’s average cost factors

for these regions, Ontario’s costs are assumed to be 16% higher for solar than the U.S. average and 10%

higher for wind. These factors are applied only to capital and fixed costs.

The regional multipliers for use with storage, nuclear, and natural gas fixed costs have been assumed to

be 1.03 based on the Leidos average.

The LCOE differences for nuclear and gas generation are summarized in Table 24 are minor and only reflect

the 3% capital premium applied.

Table 24 – Ontario vs. U.S. LCOE of Generation

Nuclear SMR CCGT w/ CCS*

U.S. 97 75 133

Ontario 98 76 135

Table 24 - Ontario vs. U.S. LCOE of Generation ($CAD/MWh)

Costs include Tx and contingency.

*49% capacity factor and including a cabon price at $100/t for

emission not captured by CCS.

Figure 92 – Electrical Regions in the U.S. Ontario’s direct

neighbors are

expensive, the rest

of the eastern U.S. is

less so


99


6.1.3 Ontario Geography Capacity Factor Implications for Renewables

Geography is important as it impacts weather patterns and subsequently the output from renewables.

Factors including how much the sun shines and how frequently the wind blows. These two factors are

geographically dependent and have a significant impact on the cost of renewable installations.

Figure 93 illustrates the number of hours of sunshine that are prevalent in various regions across North

America. 83 Ontario has far fewer hours of sunshine than the U.S., about 2,100 sunny hours per year84, the

lower end of the range for north eastern U.S. Ontario has less than half the sunshine that Arizona sees, as

indicated by the small red region in Figure 93. As a result of fewer sunny hours in a year, the average

annual capacity factor of solar generation is commensurately reduced.

Figure 94 summarizes the capacity factors of solar energy and the associated LCOE for various regions of

the U.S. in 201785. Ontario’s capacity factor for its grid-scale installations was 18% in 2017 prior to

curtailment action being taken by the IESO86. Capacity factors directly impact the cost of solar generation.

According to Lazard’s cost assumptions, the assumed capacity factor for the southwest U.S. was 28%

83 Landsberg, 1978 84 Osborne, 2018; Strapolec analysis 85 Lazard LCOE v11.0 86 IESO actuals

Hours of sunshine impacted modestly by latitude and mostly by cloud cover

> 4,000 h

3,600 – 4,000 h

3,000 – 3,600 h

2,400 – 3,000 h

2,000 – 2,400 h

1,600 – 2,000 h

South/Central Ontario: 2,100 h

Heat Map: Annual Hours of Sunshine

Figure 93 – Annual Hours of Sunshine


100


leading to solar LCOEs of $40/MWh. Ontario’s solar power will be over 50% more costly than in those

sunny U.S. states for the same installed equipment.

The Lazard capacity factor assumptions represent the conditions under which the costs have been

forecast. The LCOE for community installations assumed a capacity factor of 22.5% as discussed earlier.

Ontario’s capacity factor for community installations is assumed to be 17% based on the ratios of Lazard’s

assumptions for its grid and community scale costing scenarios87 and IESO data for Ontario’s embedded

solar generation.88 This suggests that Ontario solar could cost approximately 40% more than the average

U.S. equivalent installation.

Wind resources will be similarly affected. Figure 95 summarizes the wind capacity factors for various

regions of the U.S.89 In 2017, the capacity factor of Ontario’s wind installations was 34%90. This is

significantly lower than the Lazard LCOE capacity factor assumption of 47% for the average U.S. grid-based

wind. This means Ontario LCOEs could be 65% higher than the U.S. averages forecasted by Lazard.

87 Note that the IESO assumed that future solar would have a 15% capacity factor in its 2016 OPO 88 IESO OPO 2016 89 Lazard LCOE v11.0, U.S. average not same as LCOE chart as it is for best case scenario 90 IESO data prior to curtailment

Figure 94 – Solar (Grid-Scale) LCOE and Capacity Factor


101


6.2 Total System Costs

If the DER options were relied upon to fully provide for the province wide LTEP 2035 community daytime

demand, then the resulting architecture and capacity requirements would be as summarized in Table 25.

The architecture defines how much capacity of each generation and storage type is required to implement

the DER options. For the DER options, 16,500 MW of solar, 9,100 MW of wind, or 2,700 MW of nuclear

would be required.91

The backup gas capacity required varies from 5,700 MW for the nuclear baseload-supplied DES option to

8,500 MW in the wind-based DER option, not including reserve capacity margins. Common to all the

DER/DES options is a need for 3,000 MW of backup natural gas-fired generation to satisfy summer peak

demand. Setting aside this 3,000 MW reduces the required backup generation to 2,700 MW for the

nuclear baseload-supplied DES option, less than half the backup required by the wind-based DER option

of 5,500 MW.

91 Note that an additional 2,250 MW of baseload supply would be required to satisfy the community baseload demand.

Figure 95 – Wind (Grid-Scale) LCOE and Capacity Factor


102


Table 25 – System Architecture Assumptions for DER Options

All of these DER/DES options provide reductions in emissions from natural gas-fired generation. The

amount of natural gas-fired generation required would represent only 3 to 5% of the total supply mix. This

is equal to or less than the record 4% of the supply mix that natural gas-fired generation represented in

201792. The share of output from each generation type in the total supply mix is summarized in Table 26.

Table 26 – Supply Mix Contribution to Total Ontario Demand

Using the total system architecture requirements, Figure 96 shows the annualized cost of the DER options

which consists of three components: Generation cost, storage cost, and backup natural gas-fired

generation cost. The total cost is the sum of the annual cost of each component. The LCOE that reflects

this total cost is determined by dividing the total cost by the demand being served, which for all three

cases is 25,384 GWh as shown in Table 25.

92 IESO Year End Review 2017

DER Component Characteristic Wind Solar Nuclear

Total (GWh) 25,384 25,384 25,384

Peak (MW) 8,852 8,852 8,852

Capacity (MW) 9,116 16,535 2,698

Capacity Factor (%) 32.3% 19.1% 91.5%

Capacity (MWh) 241,275 44,375 11,240

Capacity (Hours) 62 11.5 2.9

Gas Peak (MW) 8,537 7,151 5,679

Summer Peak (MW) 3,000 3,000 3,000

Less Summer Peak (MW) 5,537 4,151 2,679

Total Required (GWh) 4,592 7,629 6,139

Table 25 - Full Year Demand Scaled-Up System Characteristics for DER Options

Community Daytime

Demand

Generation

Storage

Backup Generation

Wind Solar Nuclear

71% 71% 71%

13% 13% 13%

DER + Storage 14% 12% 13%

Gas 3% 5% 4%

Commited Resources

Community Baseload

Supply Source

Supply

Required

Table 26 - Supply Make Up of Total Ontario Demand


103


Costs for all DER/DES options are dominated by the fixed costs that reflect the annualized cost of capital

to construct and commission the solution as well as the fixed annual operating costs. Only nuclear and

gas-fired generation options have variable costs which include operations, maintenance and fuel. Nuclear

variable costs are a small portion of the total.

The cost of the nuclear generation and associated storage components is less than half the cost of the

solar and storage components of the solar-based DER solution. When excluding the common gas-fired

peaking capacity, the costs of the backup generation for the wind-based DER option is $1.9B/year or 60%

more than the $1.2B/year cost required to backup the nuclear baseload-supplied DES solution.

6.3 Intermittency and Demand Fluctuation Impacts on LCOE

As discussed in Section 5, renewables intermittency and demand fluctuations create production

inefficiencies for all generation types. Figure 97 summarizes the impact these factors have on the LCOE of

solar-based DER, wind-based DER, and nuclear baseload-supplied DES solutions93.

93 Portraying the intermittency and demand fluctuations impacts on LCOEs involves an analytical method different from that used to the total costs in Section 6.2. Wind-based DER LCOEs for the two methods differ by 3%, solar-based DER LCOEs differ by 1%.

Figure 96 – Total Annual Cost of DER and LCOE Comparison, Ontario, 2030


104


Demand fluctuations decrease the wind-based DER option costs but significantly increase the solar-based

DER and nuclear baseload-supplied DES option costs.

The inefficiencies caused by intermittency and demand fluctuation include:

• More underutilized generation

• Unused storage capacity

• Need for backup generation

The degree to which these factors impact on the LCOE varies by option.

6.3.1 LCOE for Solar-Based DER

Table 27 summarizes the performance characteristics of the solar-based DER option for the aggregated

community demand scenario. These performance characteristics impact the three cost components of

solar-based DER:

1. Cost of solar energy used to supply demand

2. Cost of storage

3. Costs of backup supply

Figure 97 – Intermittency Implications on LCOE of Ontario DER Options


105


Table 27 – Factors Impacting Solar-Based DER Costs

The factors in Table 27 contribute to the LCOE as illustrated in Figure 98. Intermittency increases the cost

of solar generation, reduces the used storage capacity, and requires backup generation.


Used Directly 979 35%

Excess 868 31%



Capacity Factor (%)


Losses 133 5%

Capacity Factor (%)

GWh % of Used GenerationBackup Generation 770 43%


Summer Peak (MW)

Less Summer Peak (MW)

Capacity Factor (%)

419

Backup

Generation

19.1%

Generation

Storage

303

Table 27 - Factors Impacting Solar-Based DER Costs

52.1%

722

21.0%

Figure 98 – Community Solar-Based DER Component LCOE Contributions


106


a) Solar cost impacts

Two factors used in this analysis impact the cost of solar energy: underutilized excess generation; and,

losses resulting from the inefficiencies of the storage device.

Excess generation impacts on the LCOE by increasing the cost of the utilized solar output. If 31% of the

solar is wasted excess, then only 69% is used. This adds a premium of $38 to the LCOE as shown in Figure

98.

Similarly, when storage losses occur, the effective amount of solar energy that can be used is reduced.

This cost specifically applies to stored solar energy, however, the net impact can be applied to the total

cost. With 5% of the total utilized solar output is lost through storage, the LCOE for solar would increase

by $6 as shown in Figure 98.

b) Cost of storage

The storage contributes to the LCOE results through two factors: the amount of utilized energy drawn

from storage; and the unused storage capacity.

The total LCOE is a function of the weighted average cost of solar utilized directly and from storage. If 60%

of the solar is utilized directly and 30% of the solar energy via storage, the new LCOE would be (60% * cost

of solar after surplus + 30% * cost of stored solar plus the nominal cost of storage). This storage adds $54

to the direct cost of solar after wastage and losses.

The impact of unused storage capacity is calculated in the same way except that the cost of storage is

increased. The cost forecast assumes that the storage is sized to produce the rated output for 8 hours per

day 350 days per year94. This is defined as the 100% capacity factor for storage. If only four hours of storage

output is available or utilized, the capacity factor will be 50%. At this capacity factor, the effective per

MWh cost of the energy drawn from storage will be double. This is because storage is a fixed cost and if

only half as much energy is extracted from the unit, the LCOE will have to double to recover that fixed

cost from half as much energy. For the solar-based DER storage capacity factor of 52%, the new weighted

average total cost of the solar-based DER system is estimated at $272/MWh, or $67/MWh more than if

the storage capacity factor was 100%.

c) Cost of Backup Generation

The estimated cost of backup generation is based on several factors: capacity factor, heat rate, and actual

MWh output.

It is assumed that 300 MW of capacity is obtained from a natural gas peaker plant for all cases. This peak

reserve capacity mostly sits idle and only operates a few days per year, when demand is especially high.

The peaking gas plant adds a fixed total cost for all of the scenarios assessed in this study.

As an example, for the solar-based DER option, the peak demand for natural-gas-fired backup generation

capacity is 722 MW. Of that, 30 MW is to be provided by the peaker plant at an LCOE of CAD $50/MWh,

which is predominantly composed of fixed costs since the peaker plant sits idle for the majority of the

time. This leaves a need for 419 MW of CCGT with CCS capacity to produce 769 GWh of output implying

94 Lazard LCOS v3.0


107


a capacity factor of 21%. The fixed portion of the CCGT w/ CCS gas plant costs US $131 per MWh. The heat

rate is 8,304 mmbtu/MWh or 10% greater than the assumed heat rate for a CCGT with CCS at an 87%

capacity factor. The variable cost for the CCGT with CCS becomes US $53/MWh, which includes fuel and

O&M plus $4/MWh of carbon pricing for the emissions that escape the CCS process (based on a carbon

price of CAD $100/t).

The total cost of backup generation becomes US $187/MWh or CAD $237/MWh including the exchange

rate, Ontario’s capital premium, and the cost of Tx and contingency.

To compute the total LCOE for the DER system and backup generation, the weighted average costs of the

directly used solar, stored solar, and backup is computed. The backup generation increases the LCOE of

the system by $6/MWh to $272/MWh as shown in Figure 98.

If a microturbine option is pursued in lieu of the centralized CCGT with CCS the costs are different. For a

microturbine with a capacity factor of 21%, the heat rate is 11,718 mmbtu/MWh with an associated

variable cost of US $72/MWh including fuel, plus US $59/MWh of carbon pricing at CAD $100/t of

emissions. Tx and contingency costs are added by scaling from EIA factors for the CCGT with CCS case, and

represent 6% of microturbine costs (the lowest Tx and contingency assumption compared to other

generation types in this analysis). The cost of 303 MW of peaker plant capacity is added as with the CCGT

with CCS scenario. The use of microturbines increases the net blended LCOE by $29/MWh to $301/MWh.

6.3.2 LCOE for Wind-Based DER

The factors impacting the costs of wind-based DER fall into the same three categories as discussed

previously for the solar-based DER case:

1. Cost of wind energy used to supply demand

2. Cost of storage


Table 28 summarizes the performance characteristics of the wind-based DER option that impact the three

cost components.


108


Table 28 – Factors Impacting Wind-Based DER Costs

The LCOE of the fully backed up wind-based DER option is summarized in Figure 99. The major contributors

are the costs of storage and the cost of backup generation. The storage and backup contributions to the

LCOE are considerably higher than for the solar-based DER case because the capacity factors of these

assets are very low, 3.5% and 9.5% respectively. The backup generation capacity factors are lower than

the solar solution because the wind option does not reduce the maximum peak demand but only requires

backup generation to output half the energy.



Excess 271 10%



Capacity Factor (%)


Losses 239 9%

Capacity Factor (%)




Summer Peak (MW)


Capacity Factor (%)

Table 28 - Factors Impacting Wind-Based DER Costs

Generation

32.3%

Storage

3.5%

862

9.5%

559

303Backup

Generation

Figure 99 – Grid-Scale Wind + CAES Component LCOE Contributions


109


The results suggest that the wind-based DER option cold perhaps be optimized further by significantly

reducing the storage capacity. A smaller storage however, would lead to greater wind output

underutilized and more gas-fire backup generation required.

6.3.3 LCOE for Nuclear Baseload-Supplied DES

The factors impacting the costs of the nuclear baseload-supplied DES fall into the same three categories

as those discussed previously for both the solar and wind-based DER cases:

1. Cost of nuclear energy used to supply demand

2. Cost of storage


Table 29 summarizes the performance characteristics of the nuclear baseload-supplied DES option that

impact the three cost components.

Table 29 – Factors Impacting Nuclear-Based DER Costs

The LCOE of the fully backed up nuclear baseload-supplied DER option is summarized in Figure 100. The

cost of the nuclear plant and the required natural gas-fired backup generation are the major contributors

to the LCOE. The backup costs appear considerably higher than for the solar case, however, that is only

because the overall nuclear baseload-supplied DES option costs are low. The LCOE of the backup

generation is $264/MWh. The backup generation costs for nuclear are lower than for solar due to the

lower backup capacity, the lower backup generation required, and the associated higher capacity factor.



Excess 221 10%

Into Storage 132 6%


Capacity Factor (%)


Losses 18 1%

Capacity Factor (%)




Summer Peak (MW)


Capacity Factor (%)

Storage

28.6%

573

26.2%

Table 29 - Factors Impacting Nuclear-Based DER Costs

Generation

91.5%

303

270

Backup

Generation


110


In the above comparison of the LCOEs of the DER options, the SMR LCOE is substituted for conventional

nuclear generation with all other assumptions remaining the same. The cost for an SMR was assumed to

be approximated CAD $75/MWh.95

6.3.4 LCOE Impact summary

Simulations of the DER/DES options using on Ontario’s demand environment and three years of wind and

solar output characteristics produced results that can be used to estimate the LCOEs of these options.

Figure 101 summarises the results that indicate that the wind-based DER option could cost $391/MWh,

solar-based DER $272/MWh, and nuclear baseload-supplied DES could be approximately $165/MWh.

Many DER concepts also advocate for gas-fired microturbines to provide local backup generation.

However, the microturbines could add another $29/MWh to the solar-based DER solution.

95 EIRP, 2017; cost of Tx and contingency added per EIA 2017

Figure 100 – Nuclear Baseload-Supplied DES Component LCOE Contributions


111


6.4 U.S. Geography Induced Competitive Disadvantages for DER

As discussed in Section 6.1.3, weather induced impacts on the output of solar and wind generation are on

average less in the U.S. As a result, the U.S. benefits from higher capacity factors than in Ontario. This

section examines how the system architecture of the DER options for the U.S. would be impacted by the

higher capacity factors and associated lower intermittency.

Figure 102 illustrates how the components of the LCOE are impacted by differences between the

jurisdictions. While solar-based DER in Ontario could be 35% higher than in the U.S., a nuclear baseload-

supplied DES option in Ontario will be 20% less expensive than U.S. installations of solar-based DER. This

suggests that Ontario may have a competitive advantage opportunity on electricity costs if that option

was pursued. This section discusses how differences in U.S. capacity factors have been used to derive the

U.S. equivalent LCOE for the solar-based DER and wind-based DER options. The three cost drivers as

discussed in Section 6.2 are: generation capacity, installed storage capacity, and installed gas-fired backup

generation capacity and its use.

Figure 101 – LCOE Comparison of DER Solutions


112


Note: in USD, solar-based DER would cost approximately $190/MWh, and wind-based DER would be about $300/MWh.

6.4.1 Solar Capacity Factor Impacts

The average solar capacity factor in the U.S. is 26% versus 19% in Ontario96. Figure 103 illustrates the

capacity factor differences by month for Ontario and the U.S.97

96 IESO actuals for 2015 to 2017 averages by solar farm, before curtailment. 97 EIA Electric Power Monthly with Data for October 2017, Table 6.7.B

Figure 103 – Solar Capacity Factor Comparison

Figure 102 – Ontario vs U.S. DER LCOE Contributions


113


The differences between the average capacity factor in Ontario and the U.S. are caused by two factors:

on average there are more hours of sunshine in the U.S. in a year; and, solar output degrades more rapidly

in Ontario in the winter than in the U.S. The smaller month-month deviations in the U.S. solar capacity

profile should improve the effectiveness of storage by reducing the seasonal variability effects.

For the purpose of modelling, a scaled U.S. profile was created to intersect and align with Ontario’s

September capacity factor, the reference month used to align the solar output with size the storage and

meet demand.

The Ontario simulations of solar-based performance indicators were analysed to identify relationships

between the capacity factor, underutilized storage capacity, and backup generation. The constant

demand scenario method described earlier was used to isolate the effects of demand fluctuations from

intermittency. The derived correlations define, for a given solar capacity factor, what the expected used

storage capacity factor will be as well as how much natural gas-fired backup generation will be required.

Implications of solar capacity factors were assessed under two scenarios: U.S. solar capacity factors; and,

the adjusted U.S. solar capacity factor. The results are summarized in Table 30 identifying the results used

to estimate the cost impacts.

Table 30 – U.S. Solar Capacity Adjustments

Since the U.S capacity factors are higher than in Ontario, less solar capacity must be installed to achieve

the same energy output, otherwise an “oversized” system would result. By reducing the installed solar

capacity matches the desired output to the required demand. To match the Ontario September design

capacity factor as shown in Figure 103, 12% less solar capacity is required. By aligning similar solar output,

the required storage capacity is unchanged.

There are limitations to this method as some of the intermittency benefits of the higher solar capacity

factors are lost. These lost intermittency benefits could increase the battery capacity factor and reduce

the need for backup generation. This is illustrated by Table 30 which contrasts the results of the U.S. and

adjusted capacity factors. To account for this limitation, and given the uncertainties inherent in this

approach, the average outcomes of the two cases was used for cost purposes.

Figure 104 illustrates the relative differences for the cost drivers of solar-based DER installations in the

U.S. and Ontario.


114


For a constant demand scenario, a U.S. capacity factor that is on average 36% higher than Ontario leads

to 16% less installed solar, 14% more solar being stored, a 25% higher battery capacity factor, and requires

68% less backup gas generation.

Figure 105 illustrates the cost implications. U.S. LCOEs can be expected to be 25% or $70/MWh less than

in Ontario.

Figure 104 – Solar Capacity Factor Implications on U.S. Cost Drivers

Figure 105 – Ontario vs. U.S. Solar-Based DER Cost Implications


115


6.4.2 Wind Capacity Factor Implications

Figure 106 illustrates the capacity factor differences by month for Ontario and the U.S.98 As with the solar-

based DER analysis, a scaled proxy to the U.S. capacity factors was created to align expected generation

output for the reference months. For wind, those months are the five highest generation fall/winter

months.

The U.S. experiences stronger winds throughout the spring and summer. The seasonal decline from winter

to summer is more moderate in the U.S., declining from a peak of 40% to a low of 25% - an overall drop

of 40%. In Ontario, the decline for the same period is from 45% to 18%, a 60% overall drop.

Correlations were established between the wind capacity factor and DER storage use and backup gas

generation need. The results are summarized in Table 31 for both the U.S. nominal capacity factor and

the adjusted capacity factor profile. As with solar, the average of the two cases has been adopted or

costing purposes.

Table 31 – U.S. Wind Capacity Adjustments

Figure 107 illustrates the relative differences in expected cost drivers for the U.S. and Ontario wind-based

DER options, highlighting the impact that a more consistent wind regime can have on the performance of

a DER system.

98 EIA Electric Power Monthly with Data for October 2017, Table 6.7.B; IESO actuals

Figure 106 – Average Monthly Wind Capacity Factor


116


The results of expected U.S. installed cost drivers are applied to the LCOE components as illustrated in

Figure 108 which shows that U.S. costs could be 10%, or $40/MWh, less than in Ontario.

6.5 Summary of Ontario’s Geographic Cost Disadvantage

Converting the costs of renewables and storage into Canadian currencies enables a comparison to existing

electricity costs in Ontario. The exchange rate, amount of domestic content, higher costs to build in

Figure 107 – Wind Capacity Factor Implications on U.S. Cost Drivers

Figure 108 – Ontario vs. U.S. Grid-Scale Wind-Based DER Cost Implications


117


Canada, and the expected capacity factors significantly increase the comparative cost of these

technologies.

The intermittency of renewables makes them uneconomic beyond 2030. When demand fluctuations are

reflected, the renewables-based DER options become 65% to 135% more expensive than the nuclear

baseload-supplied DES option.

Unfortunately, introducing more renewables in Ontario will propagate a systemic and structural energy

cost disadvantage for Ontario vis-a-vis the U.S.


118


7.0 Summary Findings and Observations

The results of this study have produced the following three major findings:

Finding #1 – Ontario’s Weather-induced intermittency undermines economics of renewables-based DER

Figure 109 summarizes the costs of the assessed DER options for meeting Ontario’s 2035 supply gap.99

The costs are compared to those of the existing system100 as well as other new developing technologies

in terms of the LCOE and the costs expected from each generation type as required to meet Ontario’s

anticipated LTEP supply gap in 2035. Intermittency creates a need for gas backup, which leads to a high

LCOE from renewables-based DER systems.

1. LCOE of solar-based DER is $270/MWh, 20% higher than the $240/MWh cost of today’s supplies that

it would replace

a. Today’s supply mix would have an LCOE of $240/MWh (if include carbon pricing).

b. Solar-based DER would have an LCOE of $270/MWh in 2017 dollars101, or $301/MWh if

microturbines are deployed instead of combined cycle gas turbines (CCGT).

99 The scenarios all assume the LTEP demand for 2035 within a system built on Ontario’s committed hydro and nuclear assets and reflect industry projected 2030 costs. 100 Existing system costs from OEB RPP, OPO 2015 embedded generation, IESO 2016 Year End data 101 All currencies in this document are in $2017 CAD except in Section 4.0 or where otherwise specified.

Figure 109 – Total Annual Cost of DER for Ontario, 2030


119


c. The LCOE of the wind-based DER option would be approximately $380/MWh.

2. A full rollout of renewables-based DER could add $0.7B/year to $3.4B/year to Ontario’s cost of

electricity.

a. This is equivalent to a cost increase of 3% to 15% over the 2030 forecast LTEP costs102.

b. The DER/DES options have two distinct cost components: the cost of generation and storage; and

the cost of the backup natural gas-fired generation and peaking supply.

c. The future generation and storage cost of a solar-based DER option is projected to be $4.8B/year,

1.9 times the cost of a baseload-supplied DES system comprised of conventional nuclear

generation and Li-ion battery storage. The generation and storage costs for the wind-based DER

option are projected to be $7.4B/year, 2.9 times the cost of the nuclear baseload-supplied DES

solution.

d. All options include the same need for peaking natural gas-fired generation plants to satisfy the

extreme demand peaks that occur on a few days every summer. The cost for 3,000 MW of peaking

gas supply in 2030 is forecast to be about $380M/year.103

Natural gas-fired generation would still be required to supply 20% to 30% of the incremental

daytime demand mostly as a result of seasonal variations in both demand and generation. The

estimated future share of natural gas-fired generation output could range from 3% to 5% of the

Ontario supply mix, similar to the 4% in 2017, but less than the 8% realized in 2016104.

The cost of the backup natural gas-fired generation required by wind-based DER is $1.9B/year,

12% more than the $1.7B/year required for solar-based DER and 62% more than the $1.2B/year

required for the nuclear baseload-supplied DES option. The wind-based DER cost is higher due to

a much greater need for backup gas-fired generation capacity.

Some proposed DER schemes involve the use of microturbines in lieu of the grid-based

generation. The incremental cost of a microturbine was examined for the solar-based DER option.

Microturbines would increase the cost by 9% due to higher capital costs, low capacity factor and

carbon pricing.

3. The impact of renewables intermittency on the LCOE of DER/DES options in Ontario is illustrated in

Figure ES-2. Intermittency results in excess unutilized generation, conversion losses in the storage

system, low capacity factors of the storage asset, and the need for backup generation.

a. The LCOE of the solar-based DER has four contributing components:

• The cost of solar panels is based on the forecast LCOE for grid-connected solar of US

$47/MWh (for low cost areas of the U.S. with high levels of sunshine105). That same

102 It is assumed that the OPO Outlook B total cost forecast of $20.2B/year in 2030 is the basis for the LTEP. 103 EIA 2017 Annual Energy Outlook, Strapolec analysis. 104 IESO Year End Data, 2016, 2017 105 Lazard LCOE v11, 2017


120


technology installed at community-scale in Ontario will cost $120/MWh, a generation

premium of $73/MWh.

• Solar output intermittency combined with demand fluctuations increases the cost of storage

and solar output by $91/MWh for the energy that is used.

• DER solutions do not eliminate surplus from intermittent renewable energy production. Up

to 30% of solar energy will be curtailed or lost through storage inefficiencies – 19% of wind.

• Natural gas will be required to backup up the solar energy and supply 30% of the demand

increasing the total LCOE by $6/MWh to $270/MWh.

b. Wind-based DER solutions are costlier at $380/MWh.

4. Residential renewables-based DER will be uneconomic for decades

To best provide the desired system asset optimization and customer benefits, DER solutions should

be located as close to the demand load as possible, preferably on the consumers’ premises.

Unfortunately, as DER systems are moved closer to loads, the scale of the DER installation decreases:

a 1.5 MW solar panel could supply a community of 1,000 homes and businesses; for a single home, a

0.25 kW solar panel could provide all of the daytime energy that would be needed above what could

be supplied by Ontario’s committed baseload. The components of 5 kW or smaller scale DER solutions

are prohibitively expensive without the substantial subsidies that have promoted their use.

a. Cost forecasts show residential solar-based DER solutions will remain uneconomic beyond

2030.

Figure 110 – Ontario vs U.S. DER LCOE Contributions


121


b. For solar-based DER, community-scale solutions may be the most promising DER option.

Increasing the size of DER installations to grid-scale solar offers little system benefits or cost

advantage.

c. Wind-based DER is only economic when using grid-scale wind, which also offers the potential

advantage of being paired with lower cost storage, such as compressed air energy storage

(CAES). However, grid-scale wind does not provide the desired DER benefit of reducing the

required capacity of the transmission and distribution systems. These must accommodate the

backup natural gas-fired generation capacity which is not reduced.

Finding #2 – Ontario renewables-based DER would have a systemic 35% higher cost structure than the U.S.

Figure ES-2 shows that the cost impacts of intermittency in Ontario are greater than in the U.S. This is

primarily due to the nature of Ontario’s geography and weather conditions, which lower the capacity

factors of the renewables. The higher capacity factors in the U.S. result from less variability or

intermittency of the renewable generation output.

1. The LCOE of the U.S. solar-based DER would be $200/MWh. The $270/MWh LCOE of the solar-

based DER in Ontario (using equivalent DER components) is 35% higher.

2. Similarly, the LCOE of wind-based DER may be 12% more in Ontario compared to the U.S.

3. Pursuing nuclear baseload-supplied DES options in Ontario could create a 20% cost advantage

over the U.S. solar-based DER options.

Finding #3 – Of the known and proven technology options, nuclear baseload-supplied DES will be the

lowest-cost option in any geography that has high renewables intermittency. Nuclear baseload-supplied

DES also has the greatest potential in Ontario to achieve the desired DER benefits of mitigating distribution

and transmission costs. Cost forecasts for other technology being developed suggest that:

1. The baseload-supplied DES would have an LCOE of $160/MWh. This option could reduce Ontario’s

annual electricity cost by over $2B.

2. Small modular reactors (SMRs) may be the lower cost solution for a broad range of jurisdictions

and locations compared to conventional nuclear;

3. Natural gas-fired generation (CCGT shown) equipped with carbon capture and sequestration (CCS)

may also be a low-cost low-carbon generation option.106 However, CCGT with CCS would not offer

the cost benefits from distribution and transmission system asset optimization and would not be

emission-free – three times more emissions than the solar-based DER and four times the

emissions of the nuclear baseload-supplied DES.

Implications

1. Renewables-based DER should not be looked to as a cost-effective solution for Ontario’s emerging

capacity gap identified in the 2017 LTEP.

• Renewables-based DER can only be justified today based on either direct subsidies or indirect

subsidies enabled by market arbitrage.

106 Assuming an operating capacity factor of 49%.


122


• Investment in residential scale renewables-based DER is uneconomic. Incentives such as net-

metering, noted in the LTEP, are an indirect subsidy that would increase the total cost of the

entire electricity system.

2. Ontario’s emerging capacity gap can be best addressed by procurement of up to 5,000 MW of new

low-emission baseload electricity supply by 2035. New baseload capacity is required in Ontario to

supply two needs:

a) To fill Ontario’s emerging capacity gap for baseload supply requires the procurement of over

2,250 MW of new low-emission baseload supply.

• These resources will be required as soon as possible after the Pickering Nuclear Generation

Station retires in 2024.

• Based on 2030 cost projections, using renewables-based DER to perform a baseload function

will cost three to four times more than new nuclear stations and will not be emissions-free

because of the requirement for backup natural gas-fired generation. With the cost projections

predicated on significant cost declines to 2030, procurements prior to 2030, such as to replace

the retiring Pickering Nuclear Generation Station, will be costlier.

b) To meet daytime demand, another 2,700 MW of low-emission baseload supply is required by

2035 in order to implement the low-cost baseload-supplied DES.

3. Given the immediate requirement for low-emission baseload generation to fill the emerging capacity

gap after 2024, planning for such procurement should begin as soon as possible to best advance a

potential Ontario competitive cost advantage with respect to the U.S.

Analysis of the transmission, distribution, and reserve capacity benefits should be conducted. It would

likely show improved relative economics of baseload-supplied DES. Such analyses could additionally

inform policy and investment decision-makers about the economics of DER/DES solutions, their ability to

help address Ontario’s emerging capacity gap, and the potential to reduce overall electricity system costs.


123


Acknowledgements

This study was conceived of and proposed by Strategic Policy Economics to fill a perceived void in the

understanding of renewables-based distributed energy resources that figure prominently in the Ontario

2017 LTEP as a solution to the forecasted supply gap.

Overview of Strategic Policy Economics

Founded by Marc Brouillette in 2012, Strategic Policy Economics helps clients address multi-

stakeholder issues stemming from technology-based innovations in policy-driven regulated

environments. The consultancy assesses strategic opportunities related to emerging innovations or

marketplace conditions and identifies approaches that will achieve positive benefits to affected

stakeholders. Strategic Policy Economics specializes in framing strategic market, science, technology

and innovation challenges for resolution, facilitating client teams in determining their alternatives,

developing business cases and business models, and negotiating multi-stakeholder public/private

agreements. Marc has worked directly with federal and provincial ministries, crown corporations and

regulators, as well as with the private sector, municipalities, and non-profit organizations.

The Strategic Policy Economics team deployed to develop this report included Marc Brouillette, Scott

Lawson, Andisheh Beiki, Marty Tzolov and Jesse Berlin.

The Strategic Policy Economics team would like to thank the Power Workers Union (PWU) and the Ontario Nuclear Advocacy Committee (ONAC) for their support in substantially funding the preparation of this report.

Strategic Policy Economics would also like to thank all the individuals who shared their views and or

reviewed and commented on draft versions of this report with particular gratitude to:

• Paul Acchione, former chair of the Ontario Society of Professional Engineers (OSPE)

• Jatin Nathwani, Professor and Ontario Research Chair in Public Policy for Sustainable Energy and

the Executive Director, Waterloo Institute for Sustainable Energy (WISE)

• Paul Newall of Newall Consulting Inc.

The Strategic Policy Economics team hopes this report provides a constructive contribution to the LTEP implementation process.


124


Appendix A – References and Bibliography

Compass Renewable Energy Consulting for Ontario Ministry of Energy, Market Analysis of Ontario’s

Renewable Energy Sector. 2017

CPI. Flexibility: The Path to Low-Carbon, Low-Cost Electricity Grids. 2017

Davis, L. Why Am I Paying $65/year for Your Solar Panels? Energy Institute at Haas. March 2018.

Retrieved from https://energyathaas.wordpress.com/2018/03/26/why-am-i-paying-65-year-for-

your-solar-panels/

Dvorak, P. Max Bögl Wind puts turbine on THE tallest tower, 178m. Blade tip to 246.5m. Windpower

Engineering. Oct, 2017. Retrieved from

https://www.windpowerengineering.com/mechanical/towers-construction/max-bogl-wind-

puts-turbine-tallest-tower-178m-blade-tip-246-5m/

EIA. Assumptions to the Annual Energy Outlook. 2017

EIA. Levelized Cost and Levelized Avoided Cost of New Generation Resources in the Annual Energy

Outlook. 2017

EIRP. What Will Advanced Nuclear Power Plants Cost? A Standardized Cost Analysis of Advanced Nuclear

Technologies in Commercial Development. 2017

Essex Energy Corporation. The Study of Energy Storage in Ontario Distribution Systems. 2017

FAO. Long -Term Budget Outlook. 2017

FAO. Nuclear Refurbishment: An Assessment of the Financial Risks of the Nuclear Refurbishment Plan.

2017

Greentech Media. First Solar Made Good on Its Promise to Beat Out Gas Peakers With Solar and

Batteries. Spector, J. 2018

GTM. U.S. Front-of-the-Meter Energy Storage System Prices 2018-2022. 2018

IESO. Energy Storage. 2016

IESO. Ontario Energy Report Q4: Electricity Prices. 2017. Retrieved from

www.ontarioenergyreport.ca/pdfs/6210_IESO_2017Q4OER_Electricity_EN_Prices.pdf

IESO. Ontario Planning Outlook. 2016

IESO. Year End Review. 2017

Inflation.eu. Historic Inflation United States – CPI Inflation. 2018. Retrieved from

http://www.inflation.eu/inflation-rates/united-states/historic-inflation/cpi-inflation-united-

states.aspx

IRENA. Electricity and Renewables: Costs and Markets to 2030. 2017


125


Joskow, P. Comparing the Costs of Intermittent and Dispatchable Electricity Generating Technologies.

2011

Landsberg, H. E. in Pinna, M. L'atmosfera e il clima, Torino, UTET. 1978

Lazard. Levelized Cost of Energy Analysis– Version 11.0. 2017

Lazard. Levelized Cost of Storage – Version 2.0. 2016

Lazard. Levelized Cost of Storage Analysis– Version 3.0. 2017

Leidos. Review of Distributed Generation and Combined Heat and Power Technology Performance and

Cost Estimates and Analytic Assumptions for the National Energy Modeling System. 2017

Maloney, P. How can Tucson Electric get Solar + Storage for 4.5¢/kWh? Utility Drive. May 2017.

Retrieved from https://www.utilitydive.com/news/how-can-tucson-electric-get-solar-storage-

for-45kwh/443715/

Massachusetts Energy Storage Initiative. State of Charge. 2016

MoE. Ontario’s Long-term Energy Plan: Delivering Fairness and Choice. 2017

Mowat Centre. Representing Consumers’ Interests in Ontario’s Energy Sector. 2017

NREL. Annual Technology Baseline. 2017. Retrieved from

http://www.nrel.gov/analysis/data_tech_baseline.html

NREL. Evaluating the Technical and Economic Performance of PV Plus Storage Power Plants. 2017

NREL. Grid-Connected Distributed Generation: Compensation Mechanism Basics. 2017

Oanda. Historic Exchange Rates. 2018. Retrieved from https://www.oanda.com/fx-for-

business/historical-rates

OEB, Regulated Pricing Plan Price Report, April 2017

OPG. Sir Adam Beck Pump Generating Station. 2018. Retrieved from https://www.opg.com/generating-

power/hydro/southwest-ontario/Pages/sir-adam-beck-pgs.aspx

Osborn, L. Sunniest Cities in Canada. Current Results. 2018. Retrieved from

https://www.currentresults.com/Weather-Extremes/Canada/sunniest-cities.php

Siemens. Dohn, R.L. The Business Case for Microgrids. 2011

Strapolec. Ontario’s Emissions and the Long-Term Energy Plan Phase 2: Meeting the Challenge. 2016

Sustainable Development Technology Canada. NRStor-Hydrostor Goderich CAES Demonstration Project.

2018. Retrieved from https://www.sdtc.ca/en/portfolio/projects/nrstor-hydrostor-goderich-

caes-demonstration-project

Toronto Hydro. Underwater Energy Storage. 2018. Retrieved from

https://www.windpowerengineering.com/mechanical/towers-construction/max-bogl-wind-

puts-turbine-tallest-tower-178m-blade-tip-246-5m/


126


Appendix B – List of Abbreviations

AC – Alternating Current

AEO – Annual Energy Outlook

AI – Artificial Intelligence

ATB – Annual Technical Baseline

BOS – Balance of System

CAD – Canadian Dollar

CAES – Compressed Air Energy Storage

CAGR – Compound Annual Growth Rate

CCGT – Combined Cycle Gas Turbine

CCS – Carbon Capture and Sequestration

CPI – Climate Policy Initiative

DC – Direct Current

DER – Distributed Energy Resource

DES – Distributed Energy Storage

Dx – Distribution

EIA – U.S. Energy Information Administration

EIRP – Energy Innovation Reform Project

EPC – Engineering, Procurement, and Construction

ESS – Energy Storage System

FAO – Financial Accountability Office of Ontario

FIT – Feed-in-Tariff

GTM – Greentech Media

GW – Gigawatt

GWh – Gigawatt Hour (one billion watts being produced for 1 hour)

HOEP – Hourly Ontario Energy Price

IESO – Independent Electricity System Operator

ICT – Information and Communication Technology

I/O – In/Out

IRENA – International Renewable Energy Agency

IRRP - Integrated Regional Resource Plans

kW - Kilowatt

kWh – Kilowatt Hour (one thousand watts being produced for 1 hour)

LCOE – Levelized Cost of Electricity

LCOS – Levelized Cost of Storage

LDC – Local Distribution Company

Li-ion – Lithium-Ion

LOLE – Loss of Load Expectation

LTEP – Long-Term Energy Plan

MoE – Ontario’s Ministry of Energy

mmBTU – Million British Thermal Units

MW – Megawatt


127


MWh – Megawatt Hour (one million watts being produced for 1 hour, enough to power ten thousand

100W light bulbs for one hour)

NREL – National Renewable Energy Laboratory

O&M – Operating and Maintenance

OEB – Ontario Energy Board

OPG – Ontario Power Generation Inc.

OPO – Ontario Planning Outlook

PCS – Power Conversion System

PNGS – Pickering Nuclear Generating Station

PPA – Power Purchase Agreement

PV – Photovoltaic

SM – Storage Module

SMR – Small Modular Reactor

Strapolec – Strategy Policy Economics

t – Tonne (1,000 kg)

TWh – Terawatt hour (one trillion watts being produced for 1 hour)

Tx – Transmission

U.S. – United States of America

USD – United States Dollar

WTO – World Trade Organization


128


Contact Information

Strategic Policy Economics

Marc Brouillette

Principal Consultant

(416) 564 - 4185

[email protected]

www.strapolec.ca

a cost and implications assessment...• ontario’s weather would put businesses at a competitive...

Documents